Error rate spike alerts
Know the moment your error rate jumps, not an hour later.
Paste this into AgentBand
Every 15 minutes, check the error rate across our production projects, and if it spikes more than 50 percent above the last hour's baseline, post an alert to #incidents with the top issues driving it and links.
A sudden jump in errors usually means something just broke. Forge watches the event rate across your projects, compares it against the recent baseline, and alerts your team the moment it spikes, with the issues behind the surge so you can act before customers do.
How it works
- 1 Compare the current error rate against the recent baseline
- 2 Detect spikes above your threshold and find the issues driving them
- 3 Post an alert with the top offenders and links
Build this agent
Connect Sentry, paste the sentence above, and let it run. You can have it working in a couple of minutes.
Get started freeCommon questions
What does the Error rate spike alerts agent do? +
A sudden jump in errors usually means something just broke. Forge watches the event rate across your projects, compares it against the recent baseline, and alerts your team the moment it spikes, with the issues behind the surge so you can act before customers do.
Which tools does this Sentry agent need? +
It uses Sentry and Slack. You connect each account once in AgentBand, and the agent acts through them. Your credentials are never exposed to the model.
How often does it run? +
Every 15 minutes. You can also run it on demand at any time, or change the schedule by chatting with it.
Do I need to write any code? +
No. You describe the agent in a sentence, connect the tools it needs, and AgentBand runs it. There is nothing to code or deploy.