Honua
// pillar 05 · simple operations

Same toolchain.
New superpower.

Reliable deployments. Simple upgrades. Easy troubleshooting. Every slow request leaves a readable OpenTelemetry trace and every login flows through the identity provider you already run — both today. In private beta: every config change as a reviewable PR, and an AI DevOps agent that triages alerts, drafts the fix, and parks it for your team to merge.

GitOps · pull requestsPRIVATE BETA
#284Update Parcels FeatureServer schema · add lot_area_acresMERGED
#283Bump zoning map style · v3 rampMERGED
#282Add OIDC role mapping for field-ops crewOPEN
#281Rotate object-storage credentialsMERGED
Identity · OIDCCONNECTED
provideroidc.example.gov
ssoenabled
last sync2 min ago
roles mappedvia IdP groups
audit logunified
Telemetry · OpenTelemetryEXAMPLE WORKLOAD
p50p95p99err raterps // illustrative · real p95s come from your deployment
// terraform · opentelemetry · ai devops

Terraform in. OpenTelemetry out. AI DevOps in between.

Ship the platform as Terraform. Watch it with OpenTelemetry. Let honua-devops walk the loop when something drifts — alerts get triaged, fixes get filed as PRs against your IaC repo, your team stays in the merge button.

terraforminfra/honua.tf
# honua.tf · ship the platform with Terraform
module "honua" {
source = "honua-io/honua/aws"
version = "~> 1.0"
 
cluster = "production"
identity = "oidc.example.gov"
storage = "s3://example-spatial-prod"
features = ["arcgis-compat", "geoserver-compat", "mcp"]
autoscaling = { min = 2, max = 24 }
 
telemetry = {
exporter = "otlp"
endpoint = "otel.example.gov:4318"
}
}
honua-devops · example loopPRIVATE BETA
03:42
⚠ otel alert · p95 spike
FeatureServer · Parcels · 87ms → 412ms
trace ab18c · span 4f2a
03:43
↪ agent · triage
Likely cause: concurrent-request autoscale lag.
cross-referenced prior incidents
03:43
↪ proposed fix · infra/honua.tf
- autoscaling = { min = 2, max = 24 }
+ autoscaling = { min = 4, max = 32 }
03:44
📝 PR filed · awaiting review
gh.example.gov/infra · "fix: bump featureserver autoscale for traffic spike"
reviewers: @sre-oncall · merge gate: green ci
governed by your OIDC roles · no bypass audit-logged · same as humans
// what your platform team already knows

The toolchain isn't new.

OperationLegacy GISHonua
Roll out a schema changeGIS console wizard. Manual on every server.PR to your GitOps repo. Reviewed, merged, rolled out.
Add a user roleBuilt-in user table. Manual sync.Map an OIDC group claim. Honua reads it.
Trace a slow requestServer logs. Open a ticket.OpenTelemetry trace. Span IDs in the response.
Rotate a credentialStop the server. Edit the file.Cloud secret manager. Honua picks it up.
Roll back a releaseBackup restore. Pray.Revert the PR. Same green pipeline.
Triage an alertPage rotation. Manual log dig at 2am.MCP-connected agent tails the trace, surfaces the likely cause, drafts the fix PR.
// next

Operate GIS
like the rest of your stack.