Expand group
Composio Platform
Previous page
Next page
Triggers
Third Party Auth
Dashboard - Operational
Dashboard
Composio For You
Toolkits
Hosted web search requests failed due to a capacity limit with the upstream search provider. Capacity has been restored and we have verified full recovery. No action is needed on your end.
This incident has been resolved. Tool executions across the impacted toolkits across all versions are now working normally.
We've narrowed down the impact to only the following toolkits and only versions strictly before 20260618_00 . Impacted toolkits: COMPOSIO_SEARCH YELP INSTACART TAVILY GEMINI SEAT_GEEK BROWSER_TOOL SERPAPI . Customers using version=latest remain unaffected.
20260618_00
COMPOSIO_SEARCH
YELP
INSTACART
TAVILY
GEMINI
SEAT_GEEK
BROWSER_TOOL
SERPAPI
version=latest
The root cause has been identified and the fix is rolling out.
We are experiencing elevated error rates for the following toolkits: COMPOSIO_SEARCH, YELP, INSTACART, TAVILY, GEMINI, SEAT_GEEK , BROWSER_TOOL and SERPAPI
We are investigating elevated errors on the MCP Servers endpoint (backend.composio.dev/v3/mcp). Some requests may fail. Sessions (Tool Router) and the rest of the v3 API are not affected. We will share an update shortly.
This incident has been resolved. Auth refresh for Linear was being rate limited by Linear, it was resolved after increasing the refresh window. End users who have multiple (more than 3-4) Linear Accounts connected with their agents may still see auth refresh failures.
We're seeing elevated failures with Linear Auth Refresh for Composio Managed Connections. Some tool calls on existing Linear connections may fail. We've traced this to rate-limiting on Linear's API and are following up with the Linear team. Other integrations are unaffected - we'll post updates here.
We have identified the issue.
We have temporarily eased its restrictions. The Cloudflare WAF remains in place. If you receive a 403 with a request id resembling something like "cle:..." or some other 3 letter region prefix please try again or reach out to us.This incident has been resolved.
We've noticed Vercel's firewall has been blocking many tool execution requests and other requests to the backend.
Ability to create new connections was down
This incident has been resolved.
CLI auth was broken and all existing CLI auth was revoked hence functionally bricking CLI
Between 10:17 and 10:23 AM PT, a subset of new connection account creations returned errors. The issue has been fixed and traffic is back to normal. Existing connections and tool execution were unaffected.
Between 2:15 and 2:38 PM PT, a subset of tool execution requests returned errors. The issue has been fixed and traffic is back to normal.
Webhook triggers are disabled
We are still investigating the fix. We will post an update in a couple of hours.
Slack triggers are currently undergoing upgrade maintenance
Between 10:49 and 11:16 AM PT, new connection account creations saw elevated errors and Slack webhook trigger deliveries were delayed. The issue has been fixed and traffic is back to normal. Existing connections continued to work throughout
We are investigation the incident
Between 12:14 and 12:21 PM PT, a subset of tool execution requests saw elevated latency. The issue has been fixed and traffic is back to normal
Apr 2026 to Jun 2026
Next
