Dawnrose
Log in Start free Menu

Product Logs and metrics

See what happened, and why.

Your agent reads the same logs and metrics you do, so it can find the bug and prove the fix.

Start free Connect your agent
webproductionObservability, last 30 minutes

CPU 38%

100%50%030 min agonow

Memory 1 GB

1 GB512030 min agonow

Requests per minute 164

220110030 min agonow

5xx responses 2

63030 min agonow
@level:error
  1. 12:04:11heap out of memory in image resize worker
  2. 12:04:12replica 1 restarted (OOM)
  3. 12:06:40heap out of memory in image resize worker
  4. 12:06:41replica 1 restarted (OOM)
  5. 12:09:02heap out of memory in image resize worker
  6. 12:09:03replica 1 restarted (OOM)
  7. 12:11:37heap out of memory in image resize worker
  8. 12:11:38replica 1 restarted (OOM)
  9. 12:14:05heap out of memory in image resize worker
  10. 12:14:06replica 1 restarted (OOM)

web is running out of memoryOn Hobby

Memory has been at 97% of its 1 GB limit for over 10 minutes. Raise its memory limit to 2 GB per replica (now 1 GB).

Sample data

One search across every service.

Filter by level, status, path or plain text, across one service or the whole project. Your agent runs the same search with search_logs.

@status:>=500 across shop / production9 lines
  1. web500POST /checkout12 ms
  2. web500POST /checkout11 ms
  3. api502GET /v1/orders/104230 s
  4. web500POST /checkout14 ms
  5. api503GET /v1/stock2 ms
  6. web500POST /checkout12 ms
  7. web500POST /cart/coupon9 ms
  8. api502GET /v1/orders/104730 s
  9. web500POST /checkout13 ms
web 6api 3Most of it: POST /checkout

Everything you need to debug.

  • Runtime logs you can filter

    Search with @level:error, @status:>=500 or plain text, across one service or the whole project.

    • @level:errorerrors only
    • @service:apione service
    • @status:>=500a JSON field, compared
    • timeout OR refusedeither word
    • -healthzleave lines out

  • HTTP logs

    Every request with its status, path and time, so slow endpoints stand out.

    1. 12:04:11GET /products20038 ms
    2. 12:04:12POST /checkout50012 ms
    3. 12:04:14GET /products/4220017 ms
    4. 12:04:19GET /20012 ms

  • Alerts

    Get a message by email, Slack, Discord or a webhook when a service crosses a threshold.

    • EmailOn
    • Slack, #opsOn
    • DiscordOff
    • WebhookOn

  • Build logs

    Every step of every build, with the deploy it belongs to, kept for 7 to 30 days depending on your plan.

  • CPU and memory

    Per service, against its limits, kept for 1 to 30 days depending on your plan.

  • Real requests and commands

    Your agent sends requests to your app and runs commands in it over the private network.

Let your agent read the logs.

Start free Read the quickstart
Paste into your terminal, with the token from your dashboard
$ claude mcp add --transport http dawnrose \
    https://app.dawnrose.ai/mcp \
    --header "Authorization: Bearer YOUR_TOKEN"