Striim Cloud MCP server
Striim Cloud MCP lets you use an AI assistant to manage your Striim Cloud account and work with applications in your services. Use it to check service health, review usage, manage access, or build and operate data pipelines through natural-language requests.
Model Context Protocol (MCP) is a standard that connects AI assistants to external tools. You register the Striim Cloud MCP endpoint with your assistant and authenticate with a Striim Cloud API token. The assistant discovers the available tools and uses them to carry out your requests.
This topic covers the account-level Striim Cloud MCP server. For self-managed deployments, see the Striim Platform MCP documentation.
Understand the scope of a request
You connect once to your Striim Cloud account and the same connection supports two types of work.
Scope | What you can do | What must be available |
|---|---|---|
Account | List, create, start, stop, resize, and delete services; review billing and activity; manage users, tokens, alerts, and supported connectivity. | Account access and permission for the requested operation. |
Service | Inspect and manage applications, work with TQL and connection profiles, view application metrics, and query live streams. | A running service with its MCP server enabled, plus access to that service. |
For a service request, name the service explicitly. Striim Cloud routes the request to that service's MCP server. You do not need to register a separate endpoint for each service or obtain a Striim Platform session token.
The tools call a Striim Cloud service an installation. For example, installation_id accepts a service name or ID. An application is a data pipeline running inside a service. A fully qualified application name includes its namespace, such as Sales.OrdersToWarehouse.
Before you begin
Make sure that:
Your MCP client supports a remote HTTP server and an Authorization header, and can reach your account URL over HTTPS.
You have a valid personal access token (PAT) or service account token for that account.
For service-level tasks, the service runs Striim 5.4.2 or later with MCP enabled. The user or service account associated with your token must have access to it. Striim calls this access federation.
Note
If your service already existed on Striim 5.4.2 or 5.4.4 before service-level MCP was enabled, restart the service to enable it. Restarting a service interrupts its running applications, so plan the restart for a suitable time.
Check your permissions
A PAT uses the permissions of its owner. Service account tokens provide administrator-level access and are intended for shared automation. MCP does not grant additional permissions.
Task | Required role or access |
|---|---|
View account and service information, metrics, billing, and activity. | Any account role. |
Manage services, Kafka, Clara AI, connectivity, and SSH tunnels. | Admin or Service Admin, within the permitted scope. |
Manage account users, roles, service accounts, other users' PATs, and account alerts. | Admin. |
Grant or remove service access. | Admin or Service Admin. The service must be running to grant access. |
Use service-level tools. | Admin, Service Admin, or Developer with access to the service. Individual operations also require the relevant service permissions. |
Invitations and account-role changes accept Admin, Developer, or Viewer.
Connect your AI assistant
Find your MCP endpoint
Use the hostname of your Striim Cloud account and append /api/v1/mcp. Replace {your-account} with the account short name shown in your console URL.
Cloud provider | MCP endpoint |
|---|---|
Google Cloud |
|
Microsoft Azure |
|
AWS |
|
For example, a Google Cloud account with the short name acme uses https://acme.striim.cloud/api/v1/mcp.
Create an API token
Use a PAT for your own assistant. For shared automation, ask an account administrator to create a service account token.
Sign in to the Striim Cloud console and open Token Management.
Select Personal Token, then Create Token. For a service account, select Service Accounts and create the token there.
Select an expiry period and complete token creation.
Copy the token when it appears and store it securely. The full value is shown only once.
If you already have a PAT, use it or regenerate it. Regenerating a token immediately invalidates the previous token, so update every client that uses it.
For token limits, expiry options, and rotation procedures, see Managing API tokens in Striim Cloud.
Register the server with your client
Configure a remote MCP server using HTTP transport, also called Streamable HTTP, with these values:
Server name:
striim-cloud, or another name you choose.Server URL: your account's MCP endpoint.
Request header:
Authorization: Bearer <your-token>.
Store the token in your client's secret storage or environment configuration. Keep credentials out of shared configuration files and source control.
Connect with Claude Code
Run the following command, replacing both placeholders. Use the complete endpoint for your cloud provider.
claude mcp add --transport http striim-cloud \ "<your-mcp-endpoint>" \ --header "Authorization: Bearer <your-token>"
Then run:
claude mcp list
In Claude Code, use /mcp to check the connection. See for client configuration options.
Connect with Cursor
Add this entry to ~/.cursor/mcp.json for your user, or merge it into an existing mcpServers object. Replace <your-mcp-endpoint> with your account endpoint.
{
"mcpServers": {
"striim-cloud": {
"url": "<your-mcp-endpoint>",
"headers": {
"Authorization": "Bearer ${env:STRIIM_CLOUD_TOKEN}"
}
}
}
}Set STRIIM_CLOUD_TOKEN in the environment used to launch Cursor, then reload the client if needed. Check the server connection in Cursor's MCP settings. See for environment variable and configuration details.
Connect with another client
Use the same endpoint and Bearer header in a client that supports remote HTTP MCP connections. A client that supports only local standard input and output (stdio) needs a compatible HTTP bridge. Follow the client's instructions for remote servers and token authentication.
Verify the connection
Ask: "Show my Striim Cloud account name and cloud provider. Do not make changes."
Confirm that the returned account is the one you intended to use.
Ask: "List my Striim Cloud services and their current states."
To test service access, ask: "In service
analytics-prod, list the available service-level tools, then list the applications I can access."
Replace analytics-prod with your service name. The first two requests verify account access. The last request also verifies that the service is running, supports MCP, and permits your identity to use service-level tools.
Work with services and applications
The examples below use an Acme account with a service named analytics-prod and an application named Sales.OrdersToWarehouse. Replace these names with your own. Include the service, application, and time range when relevant.
Check service and application health
Start with the service, then inspect the application.
Ask: "Show the state and infrastructure health of
analytics-prod."Ask: "Show CPU and memory usage for
analytics-prodover the last two hours."Ask: "In
analytics-prod, show the state, throughput, and lag ofSales.OrdersToWarehouseover the last hour. Include any reported errors."
The first two requests return service infrastructure information. The third retrieves application information from inside the service. A running service can contain a stopped or failed application, so check both when investigating a pipeline problem.
If an application has failed, ask the assistant to retrieve its reported exception and describe its components before proposing a change. Use the returned evidence to assess the cause.
To inspect lineage, ask which Oracle source tables an application captures or which files a file-based application tracks. Lineage details depend on the application's adapters.
Create a service or change its capacity
Creating a service requires a name, a solution type, and a VM size tier. The cloud provider comes from your account. If you omit the region or Striim version, Striim uses the applicable default.
Ask: "List my existing services before creating another one."
Ask: "Create a Striim Cloud service named
analytics-devwith VM size M in my account's default region. Use the default Striim version."Check the reported provisioning progress, then ask for the service state and infrastructure health. Wait until it is running before using application tools.
To change capacity, first ask which VM sizes are available for the service. Then request the size you need, for example: "Resize analytics-dev to VM size L."
Resizing changes VM capacity. Changing the node count is a separate operation available only for Striim Cloud Mission Critical services. The node-count tool accepts 1 to 10 nodes; the service configuration must support the requested count. Do not combine VM resizing and a node-count change in one tool call.
Striim requests confirmation before resizing or changing the node count. Review the target service and change before approving. After the operation completes, check the service state and the applications that should be running.
Start or stop a service
Ask: "Stop analytics-dev." Striim requests confirmation before executing the operation. The same confirmation requirement applies to starting, restarting, and deleting a service.
Stopping or restarting a service interrupts its running applications. Deleting a service removes the service. Review the exact service name and operation, then verify the resulting state in the console or through the assistant.
Build and run a pipeline
Use a running service that you can access. The source and target must meet the selected adapters' setup and connectivity requirements. A connection profile stores reusable connection settings and credentials for an adapter.
Describe the pipeline and request a plan. For example: "In
analytics-dev, prepare a pipeline from Oracle CDC to BigQuery. Show the adapters, connection profiles, and application components needed. Do not create or deploy anything yet."Ask the assistant to discover the service's tools, inspect the available adapters and their properties, and list existing connection profiles.
Reuse suitable profiles or create the required profiles before creating the application. Configure credentials through an approved secure method.
Ask the assistant to retrieve the relevant TQL grammar, prepare the TQL, and check its syntax. Review the source tables, target mappings, and transformations.
When ready, ask the assistant to execute the TQL and describe the created application. Review the results for every statement and resolve errors or disconnected components.
Deploy and start the application, then check its status, throughput, lag, and errors.
A successful syntax check confirms grammar only. Adapter properties, types, and object references are validated when the TQL executes. If only part of a request succeeds, inspect what was created before retrying.
Important
Service-level TQL can change or remove application objects. It does not receive the account-level confirmation gate. Use your client's tool approval controls and the identity's Striim permissions to control execution.
Inspect live stream data
Ask: "In analytics-prod, show the next 10 events from Sales.OrderStream, then end the query."
A query on a stream returns events that arrive while the query is active. It does not retrieve historical events. A request for 10 events can remain pending if fewer than 10 arrive. For an ongoing query, ask the assistant to stop the query when you finish so that its resources are released.
Query results can contain business data. Request only the fields and records needed for the task, and use an AI client approved for that data.
Manage account operations
Give a team member access
Inviting someone to the account and granting access to a service are separate steps.
Ask: "Invite
engineer@acme.exampleto my Striim Cloud account as a Developer."Wait for the person to accept the invitation and activate the account.
Ask: "Grant
engineer@acme.exampleaccess toanalytics-prod."Ask the assistant to verify the account membership and service access.
The service must be running when you grant access. To remove access to that service, request defederation and confirm the operation. The person remains an account member but loses access to the selected service.
For permissions within a service, ask: "In analytics-prod, show my identity, roles, and object permissions." Service tools can inspect roles and grant supported permissions when your identity has permission to do so.
Manage account membership and account roles through account tools. Direct creation, update, and activation of service users are blocked through Striim Cloud MCP.
Review usage and account activity
Ask: "Show current-period usage, quota, and remaining balance for each credit account, then show the service cost breakdown."
The cost view includes service state and last-update information. A service that is not running and has not been updated for at least seven days can be flagged as potentially idle. This flag does not establish that a running pipeline has processed no events. Review application activity before deciding that a service is unused.
For an investigation, ask: "Show who started or stopped services between October 25, 2026, 08:00 and October 25, 2026, 12:00 America/Los_Angeles."
Activity requests default to the last hour. Specify the time range and time zone when investigating a particular event. Account billing summaries describe credit usage; use the billing records for your subscription when reconciling charges.
For application or component usage, ask: "In analytics-prod, show the metering summary and usage for Sales.OrdersToWarehouse." This service-level detail is separate from the account credit balance.
Manage alerts and supporting resources
Inspect the tool's supported actions and required configuration before requesting a change.
Task | Example request and expected result |
|---|---|
Account alerts | "Show the available account alert conditions, severities, scopes, and delivery channels." Use the returned options to create a rule, then retrieve it to verify the configuration. |
Service alerts | "In |
Managed Kafka | "Show the managed Kafka options and the Kafka state for |
Clara AI | "Show the Clara AI configuration for |
Private connectivity | "List the private endpoints and Agent connectivity resources for |
SSH tunnels | "List the SSH tunnels for |
Service account tokens | "List service account tokens that expire within seven days." Plan rotation and update dependent integrations; the old token becomes invalid immediately. |
Personal access tokens | "List account PAT owners and expiry dates." Review tokens that are no longer needed before requesting revocation; revocation immediately invalidates the token. |
For account alerts, discover the supported conditions through list_options. Do not assume that an arbitrary CPU threshold or duration can be configured as an account alert. Use the appropriate service alert capability when the required condition is available there.
Private Link Services support connectivity to Striim, including supported Agent connections. Private endpoints connect Striim to external systems; endpoint groups support multiple endpoints. Business continuity endpoint groups are currently Azure only. Deleting connectivity resources can break dependent pipelines.
Review and confirm changes
Selected account operations return a pending confirmation before executing. Review the operation in your AI client and approve or reject it. A pending confirmation expires after five minutes; if it expires, request the operation again.
Operation | When Striim requires confirmation |
|---|---|
Start, stop, restart, or delete a service | Every request. |
Resize a VM or change a Mission Critical node count | Every request. |
Attach managed Kafka | When attaching. |
Manage Clara AI configuration | Requests through the AI configuration tool require confirmation. |
Rotate or revoke a service account token | Every request. Rotation invalidates the old token; revocation cannot be undone. |
Remove a user's service access | Every defederation request. |
Delete an account alert, private connectivity resource, or SSH tunnel | When deleting. |
Confirmation is not required for every change. For example, creating a service, inviting a user, changing an account role, and creating or updating an account alert can execute immediately. Client approval settings are separate from Striim's confirmation flow. Application operations routed to a service use that service's permissions and behavior.
For a request that makes several changes, verify each result. If the client times out, check the current state before submitting the request again.
Operational limits
Request rates and time ranges
The account endpoint applies these limits together:
Limit | Maximum |
|---|---|
Per token | 2 requests per second. |
Per account, across all tokens | 10 requests per second. |
Per source IP address | 100 requests per minute. |
Reduce polling and concurrent requests if you receive HTTP 429. A proxied service-level request is also subject to that service's limits.
Infrastructure monitoring retrieves 1 to 1,440 minutes per request and defaults to 60 minutes. Activity analysis defaults to 60 minutes; its relative time-range parameter accepts up to 10,080 minutes, or seven days. These are request ranges, not data-retention guarantees.
For tool calls, account monitoring and activity timestamps use RFC 3339 with a time-zone offset, such as 2026-10-25T08:00:00-07:00. Service application monitoring uses epoch milliseconds. Your assistant should use the format required by the selected tool.
Availability and supported operations
Account tasks and service tasks have different availability requirements. A stopped service can appear in account results, but its application tools are unavailable until it is running.
Service tools depend on the Striim version, service configuration, and permissions. Rediscover the tool set after an upgrade or when working with a different service.
Cloud-specific resources must be supported by the account and target service. MCP does not make an unsupported adapter or network configuration available.
Some service tools are restricted in Striim Cloud. Use account tools for membership, billing periods, credit balances, and account billing summaries.
Troubleshoot a request
Symptom | Possible cause | What to do |
|---|---|---|
HTTP 400 or an invalid-argument error | The request body or tool arguments do not match the schema. | Refresh the tool list and use the current schema. Check required values and timestamp formats. |
HTTP 401 | The Bearer token is missing, expired, revoked, or invalid. | Verify the header and token. Regenerate if needed and update the client. |
HTTP 403 | Token-account mismatch or a disabled feature flag. | Check that the token belongs to the account in the URL. Review the response's reason field. |
HTTP 429 | A token, account, IP, or service limit was reached. | Reduce request frequency and concurrency. Follow a retry interval if one is returned. |
Account requests work, but application requests fail | The service is stopped, lacks MCP support, or has not granted access to the token's identity. | Check service state, MCP availability, federation, and service permissions. |
A service tool is missing or blocked | The tool is restricted, not included in that service version, or not visible to the identity. | Discover tools for that service. Use the account equivalent where appropriate. |
A confirmation expired | More than five minutes passed before approval. | Submit the operation again and review the new confirmation. |
A stream query returns no rows | No matching events have arrived since the query started. | Check that the application is processing data. Stop the query when finished. |
TQL passes syntax checks but fails during execution | An adapter property, type, object reference, or permission is invalid. | Inspect the statement errors and adapter schema. Check existing objects before retrying. |
HTTP 500 or an internal error | A downstream request failed, for example because the MCP service is unreachable. | Check the service state. If the error persists, contact Striim Support with the correlation ID, if returned. |
When contacting Support, include the account, service, requested operation, error, and the date and time with time zone. Do not include tokens or passwords.
Tool reference for custom clients
Most users can make requests in natural language. Use this section when implementing an MCP client or inspecting the calls made by an assistant. The complete account tool reference is in Appendix A; the service-level catalog is in Appendix B.
Discover tools before calling them
Connect using the MCP client's normal initialization flow, then use tools/list to discover account tools and their input schemas. Use the returned names and schemas rather than a fixed tool count.
For work inside a service, call striim_execute_service_level_mcp with tool_name set to list_tools. Then call the required service tool through the same routing tool. The service-level names and arguments come from that service's live tool list.
The following examples are JSON-RPC tool calls within an initialized MCP connection. They are not a replacement for MCP initialization.
To list running services, call striim_list_installations with {"state_filter": "running"} as its arguments.
Discover tools for a service:
{
"jsonrpc": "2.0",
"id": 2,
"method": "tools/call",
"params": {
"name": "striim_execute_service_level_mcp",
"arguments": {
"installation_id": "analytics-prod",
"tool_name": "list_tools",
"arguments": {}
}
}
}List applications in that service by changing tool_name to striim_list_apps and supplying any arguments required by the discovered schema. An account-level striim_list_users call lists account users; the same tool name inside the service router lists users within the selected service.
Choose the tool for the task
Task | Main tools |
|---|---|
Account and services |
|
Service health and capacity |
|
Service lifecycle and scaling |
|
Usage and activity |
|
Account users and service access |
|
Account alert rules |
|
Managed Kafka and Clara AI |
|
Service account tokens |
|
Account PAT metadata |
|
For application metrics, route to striim_get_monitoring_report or striim_app_status_summary. Use striim_get_monitoring for service CPU and memory. Use striim_tql_query for metadata commands and striim_adhoc_query for live stream SELECT queries.
For private connectivity, use the relevant striim_manage_private_link_service, striim_manage_private_link_service_group, striim_manage_private_endpoint, striim_manage_private_endpoint_group, or striim_manage_private_endpoint_group_dr tool. Use striim_manage_ssh_tunnels for tunnels. Their required configuration differs by action and cloud provider; inspect the live schema before creating or changing a resource.
For a gated operation, the server returns pending_confirmation and a confirmation_id. After obtaining the user's approval, the client calls striim_approve_pending_operation with that ID; this approves and executes the operation. To cancel, call striim_reject_pending_operation. Do not treat a pending confirmation as a completed operation.
Service tools restricted in Striim Cloud
Direct service-user creation, updates, and activation are blocked through striim_create_user, striim_update_user, and striim_set_user_active. Use account invitation, role, and federation tools instead.
For billing, use account tools instead of service-level striim_list_billing_cycles, striim_get_rate_plan, and striim_get_credits. Supported service-level metering tools can still provide application or component detail where available.
Related topics
Striim MCP server service-level concepts and tools
Appendix A: Complete account tool reference
This appendix contains every individual account tool entry in the Striim Cloud MCP tool reference, including the two client confirmation tools. Account tools are called directly through your account MCP endpoint. Service tools are routed through striim_execute_service_level_mcp and are catalogued in Appendix B.
Reference conventions
All account tools use JSON-RPC 2.0. Discover the current names and input schemas with tools/list. A service name or ID is supplied through the installation parameters shown below. Required fields, conditional requirements, defaults, and allowed values are included for each documented input. A field marked No can still be required for a particular action; those conditions appear in its description.
Streaming means that progressive server-sent events are supported when the client sends Accept: text/event-stream. A confirmation response contains pending_confirmation and a confirmation_id; the client must obtain approval before calling the approval tool. Confirmations expire after five minutes.
For tools that combine read and write actions, the confirmation and impact fields apply to the actions identified in the entry. Source conflicts are identified at the affected fields rather than resolved by assumption.
Request format
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "<tool_name>",
"arguments": {
"<parameter>": "<value>"
}
}
}Account endpoint errors
HTTP status | Reason | Meaning |
|---|---|---|
400 | invalid-payload | Malformed JSON-RPC request body. |
401 | unauthorized | Missing or invalid Bearer token. |
403 | forbidden | Token-account mismatch or a disabled feature flag. |
429 | rate-limit-exceeded | IP or token rate limit exceeded. |
500 | internal-error | Downstream failure, such as an unreachable MCP service. |
Installation management
striim_list_installations
List installations with optional filtering by state, cloud provider, or solution type.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
provider_filter | string | Required: No. Allowed: aws, azure, gcp. Filter by cloud provider |
solution_filter | string | Required: No. Filter by solution type |
state_filter | string | Required: No. Allowed: all, running, stopped, failed, creating, deleting. Filter installations by state |
striim_get_installation_status
Get detailed status and information about one or all installations. Pass include_metrics=true for performance data.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Installation name, numeric ID, or all for a summary of all installations |
include_metrics | boolean | Required: No. default false. Include performance metrics and detailed information |
striim_get_health_status
Get detailed infrastructure health metrics for an installation.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Installation name or numeric ID |
striim_list_available_vms
List all available VM options for an installation. Shows VM names, display names, CPU cores, memory, and size categories. Use before striim_resize_installation.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Installation name or numeric ID to get available VMs for |
striim_create_installation
Create a new Striim installation with specified configuration. Cloud provider is automatically determined from the user's account. Suggested solution_type values: Striim Cloud, Striim Cloud Mission Critical, Validata Cloud Enterprise — any other text is best-matched against the live solution catalog. striim_version is optional; omit to use that solution's default version.
Roles: Admin, Service Admin. Streaming: Yes. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
service_name | string | Required: Yes. Name for the installation service |
solution_type | string | Required: Yes. Striim solution to install. Suggested: Striim Cloud, Striim Mission Critical, VALIDATA. Best-matched against live catalog; defaults to Striim Cloud if no match. |
vm_size | string | Required: Yes. Allowed: XS, S, M, L, XL. VM size tier |
region | string | Required: No. Deployment region (e.g. us-east-1, eastus, us-central1). Omit to use account default region. |
num_nodes | integer | Required: No. min 1; max 10. Number of nodes (STRIIMMC only) |
| string | Required: No. Optional specific version to deploy (e.g. 4.3.0). Omit to use the solution's default version. |
workload_type | string | Required: No. Allowed: standard, kafka. Workload type optimisation |
striim_manage_lifecycle
Perform lifecycle operations on installations. All actions require user confirmation.
Roles: Admin, Service Admin. Streaming: Yes. Confirmation required: Yes — ALL actions (start, stop, delete, restart) Destructive: Yes.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Target installation name or ID |
action | string | Required: Yes. Allowed: start, stop, delete, restart. Lifecycle action to perform |
target_scope | string | Required: No. Allowed: all_installations, user_installations. Scope of action |
striim_resize_installation
Vertical scaling only — change the VM size (XS/S/M/L/XL). Use striim_scale_nodes for horizontal scaling.
Roles: Admin, Service Admin. Streaming: Yes. Confirmation required: Yes. Destructive: Yes.
Source conflict: The entry describes vertical resizing only, but resize_type also lists horizontal and the input table includes target_nodes. The source does not explain how those horizontal fields apply to this tool. Use striim_scale_nodes for a node-count change and verify the current resize input schema.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Target installation |
target_size | string | Required: No. Allowed: XS, S, M, L, XL. New VM size tier |
| string | Required: No. Allowed: vertical, horizontal. Scaling direction (vertical only for this tool) |
| integer | Required: No. min 1. |
striim_scale_nodes
Horizontally scale the node count of a Striim Cloud Mission Critical (STRIIMMC) installation. Only STRIIMMC installations support horizontal scaling; all other solution types will be rejected. Do not combine with striim_resize_installation.
Roles: Admin, Service Admin. Streaming: Yes. Confirmation required: Yes. Destructive: Yes.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Target STRIIMMC installation |
node_count | integer | Required: Yes. min 1; max 10. Desired node count |
Kafka and AI clusters
striim_list_kafka_clusters
List available managed Kafka VM tiers that can be attached to an installation. Call before striim_manage_kafka attach. No parameters required.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Input: none.
striim_manage_kafka
Attach, detach, or get state of a managed Kafka cluster on an installation. Attaching Kafka requires user confirmation.
Roles: Admin, Service Admin. Streaming: Yes. Confirmation required: Yes — action=attach Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Target installation |
action | string | Required: Yes. Allowed: attach, detach, get_state. Operation to perform |
striim_manage_ai
Get, update, or delete Clara AI configuration for a service. The input table lists the action values get, upsert, and delete. Customer-managed AI requires a supported provider, model, and API key.
Roles: Admin, Service Admin. Streaming: Yes. Confirmation required: Yes. Destructive: Yes for Delete
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Target installation |
action | string | Required: Yes. Allowed: get upsert delete. Manage AI config for a service by updating the configuration. User can also delete the current configuration |
| boolean | Required: No. Set to true when user wants Hosted Striim Managed AI |
provider_name | string | Required: Yes when |
model_name | string | Required: Yes when |
api_key | string | Required: Yes when |
Private Link Service for Agent connectivity
striim_manage_private_link_service
Manage PLS for private connectivity to Striim Cloud. Actions: list, create, update, delete. Supports Private Link Service in different CSPs with VPC network/subnetwork configuration. Different CSPs use different terminology for the private service infra - Private Service Connect (GCP), Private Link Service (Azure) and VPC Endpoint Service (AWS). Confirmation required only when action=delete — this operation is irreversible and will break existing pipelines using the PLS.
Roles: Admin, Service Admin. Streaming: No. Confirmation required: Conditional — only when action=delete Destructive: Conditional (action=delete)
Parameter | Type | Requirement and description |
|---|---|---|
action | string | Required: Yes. Allowed: list, get, create, update, delete. Operation to perform |
| string | Required: No. App version tag agent-automated-infra-support to be enabled. Installation name or numeric ID — required for create, list, update, delete. Target solutions: StriimE, StriimMC |
customer_accounts | string | Required: No. Customer accounts to be validated. Subscription/ project ID that will be whitelisted on the PSI — used for create, update |
Private Link Service group for Agent connectivity
striim_manage_private_link_service_group
Manage PLS group for private connectivity to Striim Cloud with faster failover during recovery. Actions: list, create, update, delete. Supports Private Link Service Group with VPC network/subnetwork configuration. Confirmation required only when action=delete — this operation is irreversible and will break existing pipelines using the PSI.
Roles: Admin, Service Admin. Streaming: No. Confirmation required: Conditional — only when action=delete Destructive: Conditional (action=delete)
Parameter | Type | Requirement and description |
|---|---|---|
action | string | Required: Yes. Allowed: list, get, create, update, delete. Operation to perform |
| string | Required: No. App version tag agent-automated-infra-support to be enabled. Installation name or numeric ID — required for create, list, update, delete. Target solutions: StriimE, StriimMC |
customer_accounts | string | Required: No. Customer accounts to be validated. Subscription/ project ID that will be whitelisted on the PSI — used for create, update |
dr_enabled | boolean | Required: No. DR enabled flag for faster failover during recovery — used for create, update |
Private endpoint
striim_manage_private_endpoint
Manage private endpoints to connect securely to an external service. Actions: list, get, create, recreate, delete. Supports private connectivity in different CSPs with VPC network/subnetwork configuration. Confirmation required only when action=delete — this operation is irreversible and will break existing pipelines using the private endpoints.
Roles: Admin, Service Admin. Streaming: No. Confirmation required: Conditional — only when action=delete Destructive: Conditional (action=delete)
Parameter | Type | Requirement and description |
|---|---|---|
action | string | Required: Yes. Allowed: list, get, create, recreate, delete. Operation to perform |
name | string | Required: No. private endpoint name — required for create get, recreate and delete |
| string | Required: No. Installation name or numeric ID — required for create, list |
target_system | string | Required: No. label for the external system being connected to — used for the create |
pvt_ep_cfg | struct | Required: No. CSP specific cfg for the private endpoint. private endpoint config — used for create |
Private endpoint group
striim_manage_private_endpoint_group
Manage private endpoint groups to connect securely to an external service. Actions: list, get, create, recreate, delete. Supports private endpoint group for creating multiple private endpoints, mainly for connectivity to MongoDB in GCP. Confirmation required only when action=delete — this operation is irreversible and will break existing pipelines using the private endpoint groups. Note that during delete operation, dr-enabled is false in this case
Roles: Admin, Service Admin. Streaming: No. Confirmation required: Conditional — only when action=delete Destructive: Conditional (action=delete)
Parameter | Type | Requirement and description |
|---|---|---|
action | string | Required: Yes. Allowed: list, get, create, recreate, delete. Operation to perform |
name | string | Required: No. private endpoint name — required for create |
group_id | integer | Required: No. numeric Group ID – required for recreate, get, delete |
| string | Required: No. Installation name or numeric ID — required for create, list |
pvt_ep_cfg | struct | Required: No. CSP specific cfg for the private endpoint. private endpoint config — used for create |
Private endpoint group for business continuity
striim_manage_private_endpoint_group_dr
Manage private endpoint groups to connect securely to an external service with support for faster failover during recovery. Actions: list, get, create, recreate, update, delete. Supports private endpoint group with DR connectivity for SF, ADLS, generic PLS in Azure . Confirmation required only when action=delete — this operation is irreversible and will break existing pipelines using the private endpoint groups.
Roles: Admin, Service Admin. Streaming: No. Confirmation required: Conditional — only when action=delete Destructive: Conditional (action=delete)
Source conflict: The description and field notes refer to update, but the action constraint list omits update. Verify the current action schema before requesting an update.
Parameter | Type | Requirement and description |
|---|---|---|
action | string | Required: Yes. Allowed: list, get, create, recreate, delete. Operation to perform |
group_id | string | Required: No. Group name or numeric ID – required for recreate, update, get, delete |
| string | Required: No. Installation name or numeric ID — required for create, list |
target_system | string | Required: No. Target system must be supported. Target_system of private endpoints – required for create |
dr_enabled | boolean | Required: No. DR enabled flag for faster failover during recovery – required during create, update |
endpoints: name, pvt_ep_cfg | array of struct | Required: No. CSP, target specific cfg, name of private endpoint. private endpoint config, name — required for create, update |
SSH tunnels
striim_manage_ssh_tunnels
Manage SSH tunnels for database connections to Striim Cloud. Actions: list, create, update, start, stop, delete. Deletion requires confirmation and is irreversible. The source's deletion warning refers to private endpoint groups; its scope needs confirmation.
Roles: Admin, Service Admin. Streaming: No. Confirmation required: Conditional — only when action=delete Destructive: Conditional (action=delete)
Parameter | Type | Requirement and description |
|---|---|---|
action | string | Required: Yes. Allowed: list, create, update, start, stop, delete. Operation to perform |
name | string | Required: No. Name of the tunnel – required for create |
tunnel_id | string | Required: No. Tunnel name or numeric ID – required for update, start, stop, delete |
| string | Required: No. Installation name or numeric ID — required for create, list, start, stop, required for update, delete if numeric tunnel_id is not provided |
jumphost: user, ip, port | struct | Required: No. jumphost details — required for create, update |
dbhost: ip, port | struct | Required: No. dbhost details – required for create, update |
Monitoring and cost
striim_get_cost_breakdown
Get a per-installation cost view combining installation state, region, and last-update timestamps with the account billing summary. Installations not updated in 7+ days and not running are flagged as potentially idle. No parameters required.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Input: none.
striim_get_monitoring
Get infrastructure-level CPU and memory monitoring metrics for a Striim installation. Use ONLY for VM/service resource queries — NOT for application-level pipeline metrics (use striim_execute_service_level_mcp with striim_get_monitoring_report for those).
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Installation name or numeric ID |
start_time | string | Required: No. RFC3339 with TZ offset. Start time (e.g. 2026-03-30T18:00:04+05:30). If omitted, uses current time minus time_period_minutes. |
time_period_minutes | integer | Required: No. min 1; max 1440. Minutes of data to retrieve. Default: 60. If start_time provided, retrieves from that point. |
striim_analyze_activity_logs
Fetch and analyze account activity logs within a specified time range. Defaults to the last 1 hour. Analyzes logs to provide insights about account activity, actions performed, users involved, installations affected, and significant events.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
start_time | string | Required: No. RFC3339 with TZ offset. Start time (e.g. 2026-03-30T18:00:04+05:30). Defaults to current time minus time_period_minutes. |
end_time | string | Required: No. RFC3339 with TZ offset. End time. Defaults to current time. |
time_period_minutes | integer | Required: No. min 1; max 10080. Minutes of logs to retrieve (ignored if start_time provided). Default: 60. |
action_filter | string | Required: No. Filter by action type (e.g. install_created, install_deleted) |
user_filter | string | Required: No. Filter by user email or name |
Service accounts
striim_list_service_accounts
List service accounts for the current account, with optional filters on expiry, creation date, and creator.
Roles: Admin. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
expires_within_days | integer | Required: No. min 1. Filter accounts expiring within this many days (e.g. 7 for accounts expiring within a week) |
created_within_days | integer | Required: No. min 1. Filter accounts created within this many days |
created_by | string | Required: No. Filter by creator (exact or partial match) |
striim_create_service_account
Create a new service account for API automation. Returns the API token immediately — store it securely, as it cannot be retrieved later.
Roles: Admin. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
name | string | Required: Yes. Name of the service account (e.g. ci-cd-pipeline, monitoring-bot). Must be unique. |
description | string | Required: No. Optional description (e.g. Used for CI/CD automation) |
expiry_time | string | Required: No. Format: <n>h or <n>d. Token expiry (e.g. 24h, 30d). Uses system default if omitted. |
striim_rotate_service_account
Regenerates a given Service account token with a new expiry
Roles: Admin. Streaming: No. Confirmation required: Yes. Destructive: Yes.
Parameter | Type | Requirement and description |
|---|---|---|
name | string | Required: Yes. Name of the service account (e.g. ci-cd-pipeline, monitoring-bot). |
expiry_time | string | Required: No. Format: <n>h or <n>d. Token expiry (e.g. 24h, 30d). Uses system default if omitted. |
striim_revoke_service_account
Permanently revoke (delete) an existing service account. Immediately disables the account and invalidates its API token. This action cannot be undone.
Roles: Admin. Streaming: No. Confirmation required: Yes — permanently disables service account and invalidates token. Cannot be undone. Destructive: Yes.
Parameter | Type | Requirement and description |
|---|---|---|
name | string | Required: Yes. Name of the service account to revoke. The account must exist. |
Billing
striim_get_billing_summary
Retrieve the billing summary for the current account. Returns current billing period usage (billing quantity, base quota, usage percentage, overage, balance credits) and the previous period's outstanding summary — one entry per credit account. No parameters required.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Input: none.
User and access management
striim_get_account_info
Get a comprehensive summary of the current Striim account: account name, credit accounts, cloud provider (CSP) information, and the full list of account users with their emails and roles. No parameters required.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Input: none.
striim_list_users
List account users with optional filtering by role or state. Returns id, email, name, role, and state for each user.
Roles: All. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
role | string | Required: No. Allowed: Admin, Developer, Viewer. Filter by role |
state | string | Required: No. Allowed: not_verified, active, locked, deleted, expired. Filter by user state |
striim_invite_user
Invite a new user to the Striim account by email. An invitation email is sent so the user can set their password and join. Role must be one of: Admin, Developer, Viewer.
Roles: Admin. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
string | Required: Yes. Email address of the user to invite | |
role | string | Required: Yes. Allowed: Admin, Developer, Viewer. Role to assign to the invited user |
striim_update_user_role
Change the account-level role of an existing user.
Roles: Admin. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
string | Required: Yes. Email address of the user whose role to update | |
role | string | Required: Yes. Allowed: Admin, Developer, Viewer. New role for the user |
striim_federate_user
Grant an existing account user access to one or more Striim installations. The installations must be in Running state. Use striim_list_installations to find installation IDs or names.
Roles: Admin, Service Admin. Streaming: No. Confirmation required: No. Destructive: No.
Parameter | Type | Requirement and description |
|---|---|---|
string | Required: Yes. Email address of the user to federate | |
installation_ids | array | Required: Yes. Array of installation names or IDs to grant access to |
striim_defederate_user
Remove a user's access from one or more installations. The user remains in the account but can no longer log in to the specified installations. This is the reverse of striim_federate_user. Revoking access from a service will remove all access for the user on that service — requires user confirmation.
Roles: Admin, Service Admin. Streaming: No. Confirmation required: Yes. Destructive: Yes.
Parameter | Type | Requirement and description |
|---|---|---|
string | Required: Yes. Email address of the user to remove | |
installation_ids | array | Required: Yes. Array of installation names or IDs to remove the user from |
Personal access tokens (PATs)
striim_list_user_api_tokens
List all active Personal Access Tokens (PATs) for the account. Returns the owner email, creation date, and expiry date for each token. No parameters required.
Roles: Admin. Streaming: No. Confirmation required: No. Destructive: No.
Input: none.
striim_revoke_user_api_tokens
Revokes an active or expired User API token.
Roles: Admin. Streaming: No. Confirmation and impact: conflicting source entries; see the note below.
Source conflict: The detailed entry names this tool striim_revoke_user_api_tokens and marks confirmation and destructive impact as No. The same reference names it striim_revoke_user_api_token in the summary and marks confirmation as required. These names and confirmation requirements conflict. Verify the registered name through tools/list; the source does not establish which entry is correct.
Parameter | Type | Requirement and description |
|---|---|---|
string | Required: Yes. Email address of the user whose PAT needs to be revoked |
Alerts
striim_manage_alerts
Manage account-level alert rules for installation state changes, resource thresholds, and other events. Actions: list_options, list, get, create, update, toggle, delete. Use action=list_options first to discover valid scope, channel, trigger-condition, and severity IDs before creating a rule. Deletion requires user confirmation and cannot be undone.
Roles: Admin. Streaming: No. Confirmation required: Conditional — only when action=delete Destructive: Conditional (action=delete)
Parameter | Type | Requirement and description |
|---|---|---|
action | string | Required: Yes. Allowed: |
id | integer | Required: No. Alert rule ID — required for get, update, toggle, delete |
| string | Required: No. Installation name or numeric ID to filter by (list) or scope to (create). Omit for account-wide scope. |
scope_id | integer | Required: No. Alert scope ID — required for create. Use |
channel_id | integer | Required: No. Notification channel ID — required for create. Use |
trigger_condition_ids | array | Required: No. Trigger condition IDs — required for create. Use |
severity_ids | array | Required: No. Severity IDs to alert on — required for create. Use |
severity_id | integer | Required: No. Single severity ID — used for update |
email_recipients | array | Required: No. Email addresses to notify — used for create and update |
severity_filter | integer | Required: No. Optional severity filter for list |
trigger_condition_filter | string | Required: No. Optional trigger condition filter for list |
time_range | string | Required: No. Optional time range filter for list (e.g. last-7-days) |
Data plane proxy
striim_execute_service_level_mcp
Route a tool call to the MCP server embedded in the selected running service. Use for applications, TQL, monitoring, users, roles, alerts, adapters, connection profiles, ad hoc queries, and lineage. Call with tool_name=list_tools first to discover that service's tools and schemas. The token's identity must have access to the service.
Roles: Admin, Service Admin, Developer. Streaming: No. Confirmation required: No (mcp-service does not gate data-plane tools — approval is the data plane's own responsibility) Destructive: Passthrough
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. Installation ID or name whose data-plane MCP server should handle this call. Use |
| string | Required: Yes. Name of the data-plane MCP tool to invoke (e.g. |
arguments | object | Required: No. Arguments object to pass to the data-plane tool (e.g. {"app_fqn": "admin.MyApp"}). Omit or pass {} for tools that take no arguments and for |
Human-in-the-loop confirmation tools
striim_approve_pending_operation
Approves and immediately executes the underlying gated tool call in one step. Pass the confirmation_id returned by the gated tool. Confirmations expire 5 minutes after creation.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. The |
striim_reject_pending_operation
Cancels the pending operation; nothing executes. Idempotent. Pass the confirmation_id.
Parameter | Type | Requirement and description |
|---|---|---|
| string | Required: Yes. The |
Appendix B: Complete service tool catalog
Invoke these tools through striim_execute_service_level_mcp for the selected running service. Set tool_name to list_tools to obtain the service's current tool names and input schemas, then supply the chosen tool's arguments through the router. The catalog below covers every tool in the linked service-level reference. It includes restricted entries so that their account-level alternatives are clear.
The tool set depends on the service version and permissions. Account authentication continues to use the Striim Cloud Bearer token; do not substitute the direct Striim Platform authentication procedure.
TQL impact: striim_execute_tql accepts mutating statements, including STOP, UNDEPLOY, DROP, and ALTER. Calls use the identity's native service permissions; the account router does not add an approval gate to these operations.
Health and connectivity
Service tool | Purpose and documented behavior |
|---|---|
| Kubernetes liveness probe; no auth required |
| Component health |
Applications: lifecycle and inspection
Service tool | Purpose and documented behavior |
|---|---|
| List all applications visible to user |
| App metadata, status, crash/quiesce reasons |
| Lists all pipeline components inside an app (sources, targets, CQs, windows, streams, types, WActionStores) + topology warnings for disconnected streams. Use to inspect app state after execute_tql or diagnose deploy failures. |
| Create from templateId or templateDefinition + flat parameter maps |
| Deploy with optional deployment plan |
| Start or resume |
| LSN/position checkpoint state |
| get_app + last-15-min monitoring report in one call |
Applications: metadata and lineage
Service tool | Purpose and documented behavior |
|---|---|
| Oracle source lineage |
| File source/target lineage |
Monitoring
Service tool | Purpose and documented behavior |
|---|---|
| Times must be epoch ms; default last 15 min |
TQL execution (read-only metadata)
Service tool | Purpose and documented behavior |
|---|---|
| Allowlist: LIST, SHOW, DESCRIBE, USE. SELECT is blocked — use |
TQL knowledge tools (AI-driven application development)
Service tool | Purpose and documented behavior |
|---|---|
| List all available source/target/parser/formatter adapters. Filter by category or name substring. Use to verify an adapter exists before writing CREATE SOURCE TQL. |
| Full property schema for a named adapter — required vs optional fields, types, defaults, tql_usage_hint. Essential before writing CREATE SOURCE / CREATE TARGET TQL. |
| TQL grammar rules, syntax summary, examples, and streaming-specific gotchas for a given statement type. Pass a single statement_type or an array statement_types to batch-fetch multiple contexts in one call. Covers: CREATE_APPLICATION, CREATE_SOURCE, CREATE_TARGET, CREATE_CQ, CREATE_WINDOW, CREATE_STREAM, CREATE_TYPE, CREATE_WACTIONSTORE, SELECT, BUILTIN_FUNCTIONS, CREATE_CONNECTIONPROFILE. Each entry includes prerequisites listing companion tools to call first. |
| Check TQL for syntax/grammar errors without executing — parser-level only, no MDR writes. Returns success, statements_parsed, a structured error with position info, and topology_warnings (streams produced but not consumed, or consumed but not produced). NOTE: syntax check only — does NOT validate adapter names, property values, types, or stream/type references; those are validated by |
| Execute mutating TQL: CREATE APPLICATION, CREATE SOURCE/TARGET/CQ/WINDOW/TYPE/STREAM, ALTER APPLICATION, END APPLICATION, DEPLOY, START, UNDEPLOY, DROP. Auto-batches large CREATE APPLICATION…END APPLICATION blocks into chunks ≤ 15 inner statements to avoid streaming response limits. Returns per-statement status array. Optionally run |
Ad hoc query (live stream inspection)
A stream SELECT sees events arriving while the session is active. A WActionStore SELECT returns a point-in-time snapshot. Use LIMIT for a bounded query; otherwise fetch successive batches using the returned next_cursor, then stop the query. Up to three concurrent sessions per principal per transport are documented.
Fetch uses an explicit cursor and does not consume the retained rows. Reusing the same cursor returns the same rows while they remain retained. The documented buffer holds up to 50,000 rows with drop_oldest as the default overflow policy. Handle cursor_expired and cursor_out_of_range as reported by the tool.
Service tool | Purpose and documented behavior |
|---|---|
| Start an adhoc SELECT against a live stream or WActionStore. With LIMIT N: blocks until N rows arrive and returns status=completed (truncated=true if rows were dropped under load). Without LIMIT: returns immediately with a query_id, the first batch, and a |
| Non-destructive cursor read. Pass cursor = the |
| Stop a streaming adhoc session and release server resources. Idempotent — safe to call on an already-stopped session. Returns total rows returned (acked high-water), rows dropped, elapsed ms. |
Connection profiles
Create connection profiles before the CREATE APPLICATION block. Discover the supported profile adapters and properties, then reuse or create the required profile and reference its fully qualified name.
Service tool | Purpose and documented behavior |
|---|---|
| List all connection profiles on the cluster — name, adapterType (connectionprofile category), namespace. Use before writing TQL to discover existing profiles to reuse, or after |
| Create or replace a connection profile via TQL. Accepts profile_name (FQN), adapter_type (from |
TQL files
Service tool | Purpose and documented behavior |
|---|---|
| List uploaded TQL files |
Users
Service tool | Purpose and documented behavior |
|---|---|
| All users |
| One user |
| Create a service user. Cloud restriction: Restricted in Striim Cloud. Use account invitation and federation tools. |
| Update service-user email, password, or properties. Cloud restriction: Restricted in Striim Cloud. Manage the user through account tools. |
| Activate or deactivate a service user. Cloud restriction: Restricted in Striim Cloud. Manage user access through account tools. |
Roles
Service tool | Purpose and documented behavior |
|---|---|
| All roles |
| One role |
| Create a role using its namespace and role name. |
Permissions (grants only)
Service tool | Purpose and documented behavior |
|---|---|
| Assign roles |
| Grant object perms |
| Role-to-role inheritance |
| Grant perms to a role |
System alerts
Service tool | Purpose and documented behavior |
|---|---|
| List system alerts for the service. |
| Retrieve a named system alert. |
| Update a system alert by submitting the full alert object. |
Metering and billing
Service tool | Purpose and documented behavior |
|---|---|
| Usage summary + current-cycle usage; each sub-call independently guarded |
| List service billing cycles. Cloud restriction: Restricted in Striim Cloud. Use the account billing summary. |
| Retrieve application and current-period usage detail. |
| Per-view required params |
| List metered entities for the requested kind. |
| Retrieve adapter or event rate-plan information. Cloud restriction: Restricted in Striim Cloud. Use account billing information. |
| Retrieve awarded-credit and depletion information. Cloud restriction: Restricted in Striim Cloud. Use the account billing summary. |
Admin / introspection
Service tool | Purpose and documented behavior |
|---|---|
| Authenticated user, roles, permissions |
| Cluster name + member servers |