- Home
- Skills
- DevOps & Cloud
- Investigate Production Issues with Flare (CLI + MCP)
Investigate Production Issues with Flare (CLI + MCP)
Triage errors, slow requests and firing alerts in a self-hosted Flare instance, moving from alert to logs to trace to root cause with the flare CLI or MCP tools, read-only.
Free
Works with the AI Agents you already use
Investigate Production Issues with Flare (CLI + MCP)
Example session with this skill installed
Checkout is throwing 500s since about an hour ago. Find out why using Flare.
- Read your context and instructions
- Compiled the investigate production issues
Cause: payments timeouts after a deploy, not checkout itself. Evidence: 41 of 46 error traces in the last hour contain a payments client span at ~30s with HttpRequestException (trace 3f9a…); payments exception count went from 0 to 52 at 14:07; checkout logs show "payment declined: upstream timeout" with the same trace ids. Suggested next step: look at the payments deploy at 14:05. Nothing was changed in Flare.
Connects securely to your tools. The creator never sees your data.
About this skill
Give your AI agent a disciplined way to find the cause of a production problem in a self-hosted Flare instance (an OpenTelemetry-native platform for logs, traces, metrics, exceptions and alerts). It uses Flare's read-only MCP tools when they're connected, and the flare CLI otherwise.
What your agent will do
- Start from the symptom: a firing alert, an error spike, a slow endpoint, or "something changed after the deploy".
- Pull the right evidence: grouped exceptions, failing or slow traces, error logs, metric step changes, unhealthy cron/job monitors.
- Pivot from a log line to its trace to the span that owns the time or the error, then back to every log from that request.
- Compare before and after a fix or restart (
compare_runs,diff_traces) to confirm a bug is actually gone. - Build an incident bundle (trace + correlated logs + metrics) for a bug report.
- Report cause, evidence, blast radius, what was ruled out, and gaps, and never change alerts or config without being asked.
Good to know
- Free and read-only by design.
- Needs a Flare instance with data. If nothing shows up, the skill checks ingestion health and
flare doctorbefore drawing conclusions. - CLI flags were checked against the flare CLI's own
--helpoutput.
Read-only investigation: it does not change alert rules, silence alerts or edit Flare configuration unless you explicitly ask. It can only see what has been ingested into Flare and is still within retention. Command flags were checked against the flare CLI's own --help output but not run against a live instance with production data.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 14 days ago
- Free with an Agensi account
Requires