Always-on cloud ops agents that detect configuration drift, self-heal incidents around the clock, and ship a curated weekly reliability digest to your engineering team — without locking you to a single cloud or charging enterprise AIOps prices.
Built by Fred Bitenyo — Senior DevOps Engineer, former pager.
mendhelm · /ops/liveCapabilities
Mendhelm is one product, four operators. Each one runs the same way your platform team would — diff first, dry-run, promote, audit — except it doesn't sleep, doesn't take holidays, and never goes a week without reporting back.
CND-01Drift detectionContinuous diff against declared and provisioned state. Catches config drift, IAM policy drift, label drift, and tag drift before they cascade into incidents.
CND-02Self-healing remediationAgents open, triage, and close incidents autonomously — restart the stuck deployment, roll back the bad release, drain the unhealthy node, raise the limit, re-attach the volume.
CND-03Patching and config rolloutPush pinned versions, rotate secrets, apply Terraform deltas, patch base images. Each rollout is sandboxed, then promoted — never the other way around.
CND-04Weekly reliability digestA curated engineering-team report every Monday: what drifted, what healed, what an on-call human should still sign off on. Not a dashboard. A summary a director can act on.
How it works
Implementation is a single OIDC connector and a one-page scope config. There is no agent to install, no sidecar to deploy, no "observability tax" to pay.
Read the agent loop deep dive →step / 01✓ readyRead-only by default. Agents assume least-privilege scopes per environment, per service, per region. You grant exactly the blast radius you trust us with.
step / 02✓ readyMendhelm baselines your environments over the first 72 hours — normal traffic shapes, deployment cadences, idle baselines. After baseline, drift becomes signal instead of noise.
step / 03✓ readyAgents remediate with dry-runs first, then promote. Every action is logged with prompt, scope, diff, and outcome. Weekly digests roll the work up for the team.
Security & controls
Giving an agent permission to change production is a trust motion. We earn it by making the controls explicit, the audit trail complete, and the off-switch always one click away.
Request the security briefaudit.last_24h · 247 actions · 0 escalations · 0 anomalous scopes
Weekly reliability digest
Every Monday, Mendhelm sends a curated engineering-team report — not a raw dashboard. What drifted. What healed. What an on-call human still owns. Reliability framed as a product engineering output your leadership and rotation can act on.
monday · 09:00 local · your inbox
Reliability digest · week 34
Mendhelm · weekly
drifted · cleared
The weekly baseline turned up six drifted assume-role policies on themendhelm-prodaccount. Each was diffed against the last approved manifest, dry-run, then promoted. No engineer was paged.
healed · self
A 7-pod crash-loop in checkout-prod cleared after a rolling restart and a softened readiness probe that won't mask a real outage next time.
escalated · human
Auth spends running 14% ahead of the rolling forecast. Mendhelm refuses to scale down further (you asked us not to) — flagging for the platform owner.
Talk to Fred
Mendhelm is in private access. Founder Fred Bitenyo takes the first call personally — what you run, what hurts, whether agents are even right for you yet. No SDR.
response within 24h · no automation · no SDR