d5s

Engineering AI agent

Monitor AWS costs and Kubernetes rollouts. Get savings suggestions and post-release findings, with the data behind each one.

Works with
Add to your team

What you get

  • AWS cost and savings signals

    Compare cost and usage, then suggest optimizations. Show data freshness and access gaps.

  • Kubernetes after a release

    Check rollout state, pod events, logs and metrics. Flag changes that need review.

  • Errors and changes in one report

    GitHub changes, Sentry errors and Grafana metrics, with the time of each read.

Daily wakeup · Review changes since yesterdayMajor release · Request an extra check
d5s · Engineering investigation
API service · production EULive signals read at 09:14
Investigate

Incoming signals

AWS health
Grafana metrics
Sentry errors
Kubernetes logs
5xx error rate2.8%
↑ since 09:12
08:4509:14
K
Cluster event

Two restarts in api-prod after 09:12

L
Pod logs

Error entries begin after the deploy

Observed signals · cause not yet confirmed

Next step · needs review

Review the deploy and pod logs with the on-call engineer.

Monitoring and observability

See what happened.
Then decide what to do.

Schedule a daily wakeup to check GitHub changes, AWS costs, Grafana metrics, Sentry errors and Kubernetes rollout health. After a major release, request an extra review. Each report shows when it read the data and what it could not access. Scheduled AWS reads need a separate grant; other sources use their own permissions.

d5s · Cluster · production EU
Morning health checkAuthorized read · 09:14
4 nodes checked
Daily wakeup · 09:00

Check GitHub changes and new signals since the last run.

Nodes4 / 4Healthy
CPU load62%
Warnings2Review
api-prod · recent restartsReview
Grafana · last read 09:14Available

How it works with your team.

Grant the workspace access to its AWS account, cluster and observability sources. Scheduled AWS reads need a separate grant. Kubernetes scheduled runs use the agent’s existing access. See all engineering jobs in AI for engineering teams.

Tools and schedule

  • AWSApproved cost and usage reads
  • KubernetesAuthorized rollout, pod and event reads
  • ObservabilityGranted logs, metrics and dashboards

Daily change, cost and health checks when scheduled; AWS reads need a separate grant

Set it up in three steps.

  1. 01

    Create the agent

    Add an Engineering agent and choose the reports your team needs.

  2. 02

    Grant access

    Connect an AWS account and cluster, then grant access to GitHub, Grafana and Sentry as needed.

  3. 03

    Test, then schedule

    Verify one live read, then schedule a daily check. Request an extra review after a major release.

Add an Engineering agent to your team

It investigates and prepares evidence. Your engineers decide what changes.

Add to your team