Elgato_dark/datahaven
1
0
Fork 0
mirror of https://github.com/datahaven-xyz/datahaven synced 2026-05-24 09:50:01 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'main' into sde/upd-noteworthiness-label

Browse source
This commit is contained in:
Ahmad Kaouk 2026-03-13 14:20:36 +01:00 committed by GitHub
commit d353d1aee0
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 298 additions and 27 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
.github/workflows/release.yml vendored
View file

@ -18,7 +18,7 @@ on:
permissions:
contents: read
packages: write # Required for docker-build-release
packages: write # Required for docker build release jobs
concurrency:
group: release-${{ github.workflow }}-${{ github.ref }}
@ -30,3 +30,9 @@ jobs:
secrets:
DOCKERHUB_USERNAME: ${{ secrets.DOCKERHUB_USERNAME }}
DOCKERHUB_TOKEN: ${{ secrets.DOCKERHUB_TOKEN }}
validator-set-submitter-docker-build-release:
uses: ./.github/workflows/task-docker-release-validator-set-submitter.yml
secrets:
DOCKERHUB_USERNAME: ${{ secrets.DOCKERHUB_USERNAME }}
DOCKERHUB_TOKEN: ${{ secrets.DOCKERHUB_TOKEN }}

105
.github/workflows/task-docker-release-validator-set-submitter.yml vendored Normal file
View file

@ -0,0 +1,105 @@
name: Docker Build & Publish Validator Set Submitter (Release)
on:
workflow_dispatch:
inputs:
label:
description: "Label for the Docker image"
required: true
type: string
branch:
description: "Branch to checkout and build"
required: true
type: string
workflow_call:
secrets:
DOCKERHUB_USERNAME:
description: "Docker Hub username"
required: true
DOCKERHUB_TOKEN:
description: "Docker Hub access token"
required: true
outputs:
image-tag:
description: "The tag portion of the docker image (without registry)"
value: "${{ jobs.build-test-push.outputs.image-tag }}"
permissions:
contents: read
packages: write
concurrency:
group: docker-build-release-validator-set-submitter-${{ github.ref }}
cancel-in-progress: true
jobs:
build-test-push:
runs-on: ubuntu-latest
# Require approval before publishing to Docker Hub
environment: production
outputs:
image-tag: ${{ steps.extract_tag.outputs.image-tag }}
steps:
- name: Checkout repository
uses: actions/checkout@v4
with:
ref: ${{ github.event.inputs.branch || github.ref }}
- uses: ./.github/workflows/actions/cleanup-runner
# --- Docker metadata ---
- name: Docker meta (dispatch)
if: github.event_name == 'workflow_dispatch'
id: meta-dispatch
uses: docker/metadata-action@v5
with:
images: datahavenxyz/validator-set-submitter
flavor: |
latest=false
tags: |
type=raw,value=${{ github.event.inputs.label }}
- name: Docker meta (CI - main push)
if: github.event_name != 'workflow_dispatch'
id: meta-ci
uses: docker/metadata-action@v5
with:
images: datahavenxyz/validator-set-submitter
flavor: |
latest=true
tags: |
type=raw,value=latest
type=sha,format=short,prefix=sha-
- name: Extract tag for job output
id: extract_tag
run: |
if [ "${{ github.event_name }}" == "workflow_dispatch" ]; then
FULL_TAG=$(echo '${{ steps.meta-dispatch.outputs.json }}' | jq -r '.tags[-1]')
else
FULL_TAG=$(echo '${{ steps.meta-ci.outputs.json }}' | jq -r '.tags[-1]')
fi
TAG_ONLY=$(echo "$FULL_TAG" | sed 's|.*:||')
echo "image-tag=$TAG_ONLY" >> $GITHUB_OUTPUT
echo "image-name=datahavenxyz/validator-set-submitter:$TAG_ONLY" >> $GITHUB_OUTPUT
# --- Build and push Docker image ---
- name: Build and push Docker image
uses: ./.github/workflow-templates/publish-docker
with:
dockerfile: ./test/tools/validator-set-submitter/Dockerfile
context: .
registry: docker.io
registry_username: ${{ secrets.DOCKERHUB_USERNAME }}
registry_password: ${{ secrets.DOCKERHUB_TOKEN }}
image_tags: ${{ steps.meta-dispatch.outputs.tags || steps.meta-ci.outputs.tags }}
image_title: "Validator Set Submitter - Release"
image_description: "Release build of DataHaven validator set submitter"
cache_scope: validator-set-submitter-release-build
# --- Smoke tests ---
- name: Pull and test submitter --help
run: |
docker pull ${{ steps.extract_tag.outputs.image-name }}
docker run --rm ${{ steps.extract_tag.outputs.image-name }} --help

212
test/tools/validator-set-submitter/README.md
View file

@ -1,6 +1,6 @@
# Validator Set Submitter
Long-running daemon that automatically submits validator-set updates from Ethereum to DataHaven each era via Snowbridge.
Daemon process that automatically submits validator-set updates from Ethereum to DataHaven each era via Snowbridge.
## How it works
@ -11,7 +11,14 @@ The submitter subscribes to finalized `Session.CurrentIndex` changes on DataHave
3. Is `ExternalIndex` already at or past `targetEra`?
4. Is the current session the last session of the era?
If all preconditions are met, it calls `sendNewValidatorSetForEra` on the ServiceManager contract. Each era gets a single submission attempt — if it fails, the era is missed and the submitter moves on to the next.
If all preconditions are met, it calls `sendNewValidatorSetForEra` on the ServiceManager contract. Submission attempt tracking is in-memory, so each era gets a single submission attempt per process run. If an attempt fails, that era is marked missed for this run and the submitter moves on to the next era.
### Runtime and restart behavior
- The submitter does not implement automatic reconnect/backoff for DataHaven session-subscription failures.
- On a subscription error, it logs the error, stops the watcher, and the process exits.
- Run it under a restart policy (for example `systemd` with `Restart=always` or Kubernetes with `restartPolicy: Always`).
- After a restart, a previously failed era may be attempted again if `ExternalIndex` has not advanced past that target era.
## Prerequisites
@ -44,6 +51,47 @@ relayer_fee: "0.2"
# metrics_port: 8080
```
### Settings reference
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
| `ethereum_rpc_url` | string | Yes | — | Ethereum JSON-RPC endpoint |
| `datahaven_ws_url` | string | Yes | — | DataHaven WebSocket endpoint |
| `submitter_private_key` | hex string | No\* | — | Private key of the authorized submitter account (`0x` + 64 hex chars) |
| `network_id` | string | No | `"anvil"` | Network ID used to locate `contracts/deployments/{network_id}.json` |
| `service_manager_address` | hex address | No\*\* | — | ServiceManager contract address |
| `execution_fee` | string (ETH) | No | `"0.1"` | Snowbridge execution fee sent as `msg.value` |
| `relayer_fee` | string (ETH) | No | `"0.2"` | Snowbridge relayer fee sent as `msg.value` |
| `metrics_port` | integer | No | `8080` | Prometheus metrics server port (1–65535) |
\* Required via one of: `--submitter-private-key` flag, `SUBMITTER_PRIVATE_KEY` env var, or `submitter_private_key` in config.
\*\* Required when running in Docker (deployment files are not included in the image). When omitted, the address is read from `contracts/deployments/{network_id}.json`.
### Private key precedence
The submitter private key is resolved in this order (first wins):
1. `--submitter-private-key` CLI flag
2. `SUBMITTER_PRIVATE_KEY` environment variable
3. `submitter_private_key` in the config YAML file
### Environment variables
| Variable | Description |
|---|---|
| `SUBMITTER_PRIVATE_KEY` | Submitter private key (see precedence above) |
| `METRICS_PORT` | Override metrics port (takes precedence over config file, but CLI flag wins) |
| `LOG_LEVEL` | Log verbosity: `debug`, `info` (default), `warn`, `error` |
### CLI flags
| Flag | Description |
|---|---|
| `--config <path>` | Path to YAML config file (default: `./tools/validator-set-submitter/config.yml`) |
| `--submitter-private-key <key>` | Override submitter private key |
| `--metrics-port <port>` | Override metrics server port |
| `--dry-run` | Log what would be submitted without sending transactions |
## Usage
From the `test/` directory:
@ -65,54 +113,127 @@ bun tools/validator-set-submitter/main.ts run --submitter-private-key 0x...
bun tools/validator-set-submitter/main.ts run --dry-run
```
Private key precedence is: `--submitter-private-key` > `SUBMITTER_PRIVATE_KEY` > `submitter_private_key` in config file.
## Observability
The submitter exposes a Prometheus metrics server on `metrics_port` (default `8080`):
The submitter exposes an HTTP server on `metrics_port` (default `8080`) with three endpoints:
- `GET /metrics` — Prometheus metrics
- `GET /healthz` — liveness
- `GET /readyz` — readiness (`200` once startup checks pass and watcher is running)
| Endpoint | Purpose | Codes |
|---|---|---|
| `GET /metrics` | Prometheus metrics scrape | `200` |
| `GET /healthz` | Liveness probe | `200` always |
| `GET /readyz` | Readiness probe | `200` when startup checks passed and watcher is running, `503` otherwise |
Key metrics:
### Metrics reference
- `validator_set_submitter_submissions_total{outcome="success|failed|dry_run"}`
- `validator_set_submitter_ticks_total{result="submitted_success|submitted_failed|skipped_*"}`
- `validator_set_submitter_errors_total{type="tick_error|subscription_error"}`
- `validator_set_submitter_missed_eras_total`
- `validator_set_submitter_consecutive_missed_eras`
- `validator_set_submitter_up`
- `validator_set_submitter_ready`
All metrics are prefixed with `validator_set_submitter_`.
#### Counters
| Metric | Labels | Description |
|---|---|---|
| `submissions_total` | `outcome`: `success`, `failed`, `dry_run` | Total submission attempts by result |
| `ticks_total` | `result`: `submitted_success`, `submitted_failed`, `skipped_no_active_era`, `skipped_already_submitted`, `skipped_already_confirmed`, `skipped_not_last_session` | Tick evaluation outcomes |
| `errors_total` | `type`: `tick_error`, `subscription_error` | Non-submission errors |
| `missed_eras_total` | — | Total eras where the submission attempt failed |
#### Gauges
| Metric | Description |
|---|---|
| `active_era` | Current active era on DataHaven |
| `target_era` | Target era for next submission (`active_era + 1`) |
| `external_index` | Latest confirmed era on-chain |
| `current_session` | Current session number |
| `last_submitted_era` | Last era successfully submitted |
| `consecutive_missed_eras` | Consecutive missed eras (resets to 0 on success) |
| `up` | `1` if watcher is running, `0` if stopped |
| `ready` | `1` if startup checks passed and watcher running, `0` otherwise |
#### Histograms
| Metric | Buckets | Description |
|---|---|---|
| `submission_duration_seconds` | 1, 5, 10, 30, 60, 120, 300 | Time from transaction send to receipt |
| `tick_duration_seconds` | 0.1, 0.5, 1, 2, 5, 10, 30 | Time to process one tick |
### Alerting recommendations
Example Prometheus alert rules for common failure modes:
```yaml
groups:
- name: validator-set-submitter
rules:
- alert: SubmitterDown
expr: validator_set_submitter_up == 0
for: 2m
labels:
severity: critical
annotations:
summary: "Validator set submitter is down"
- alert: ConsecutiveMissedEras
expr: validator_set_submitter_consecutive_missed_eras > 0
for: 0m
labels:
severity: critical
annotations:
summary: "Submitter has missed {{ $value }} consecutive era(s)"
- alert: SubmissionErrorsIncreasing
expr: rate(validator_set_submitter_errors_total[5m]) > 0
for: 5m
labels:
severity: warning
annotations:
summary: "Submitter errors increasing (type={{ $labels.type }})"
- alert: SlowSubmissions
expr: histogram_quantile(0.95, rate(validator_set_submitter_submission_duration_seconds_bucket[15m])) > 120
for: 5m
labels:
severity: warning
annotations:
summary: "95th percentile submission duration exceeds 120s"
```
## Docker
Build the image from the repository root:
A pre-built image is published to Docker Hub on every push to `main`:
```bash
docker build -f test/tools/validator-set-submitter/Dockerfile \
-t datahavenxyz/validator-set-submitter:local .
```
datahavenxyz/validator-set-submitter:latest
datahavenxyz/validator-set-submitter:sha-<short>
```
Run the submitter with mounted config and env private key:
Run the submitter with a mounted config and private key:
```bash
docker run --rm \
-v "$(pwd)/test/tools/validator-set-submitter/config.yml:/config/config.yml:ro" \
-v "$(pwd)/config.yml:/config/config.yml:ro" \
-e SUBMITTER_PRIVATE_KEY=0x... \
datahavenxyz/validator-set-submitter:local
datahavenxyz/validator-set-submitter:latest
```
Dry run:
```bash
docker run --rm \
-v "$(pwd)/test/tools/validator-set-submitter/config.yml:/config/config.yml:ro" \
-v "$(pwd)/config.yml:/config/config.yml:ro" \
-e SUBMITTER_PRIVATE_KEY=0x... \
datahavenxyz/validator-set-submitter:local --dry-run
datahavenxyz/validator-set-submitter:latest --dry-run
```
The Docker image does not include `contracts/deployments/*.json`. In containerized runs, set `service_manager_address` in your config.
The Docker image does not include `contracts/deployments/*.json`. Set `service_manager_address` explicitly in your config.
### Building locally
To build the image from the repository root:
```bash
docker build -f test/tools/validator-set-submitter/Dockerfile \
-t datahavenxyz/validator-set-submitter:local .
```
## Startup checks
@ -127,3 +248,42 @@ If any check fails, the process exits immediately.
## Shutdown
Send `SIGINT` (Ctrl+C) or `SIGTERM`. The submitter unsubscribes from session changes and tears down connections cleanly.
## Troubleshooting
### Startup exits immediately
| Symptom | Cause | Fix |
|---|---|---|
| `Cannot connect to Ethereum RPC` | Ethereum endpoint unreachable | Verify `ethereum_rpc_url` is correct and the node is running |
| `Cannot connect to DataHaven WS` | DataHaven endpoint unreachable | Verify `datahaven_ws_url` is correct and the node accepts WebSocket connections |
| `Account 0x... is not the authorized submitter` | Private key does not match the on-chain submitter | Call `setValidatorSetSubmitter` on the ServiceManager with the correct address, or fix the private key |
| `Missing submitter private key` | No key provided | Supply via `--submitter-private-key`, `SUBMITTER_PRIVATE_KEY` env var, or `submitter_private_key` in config |
| `Config file not found` | Wrong `--config` path | Check the path and ensure the file exists |
### Missed eras
When the submitter fails to submit for an era, `missed_eras_total` increments and `consecutive_missed_eras` increases. Common causes:
- **Transaction reverted** — the submitter account may have insufficient ETH to cover `execution_fee + relayer_fee`. Fund the account.
- **RPC timeout** — the Ethereum RPC may be overloaded or unreachable. Check RPC health and consider a dedicated endpoint.
- **Snowbridge congestion** — if the bridge queue is full, submissions may fail. Check Snowbridge relayer status.
- **Already confirmed** — if another process submitted the era, the submitter skips it (this is normal, not an error).
Check `LOG_LEVEL=debug` output for detailed tick-by-tick reasoning.
### Process exits after running for a while
| Symptom | Cause | Fix |
|---|---|---|
| `Session subscription error: ...` followed by process exit | DataHaven WebSocket subscription dropped and the submitter has no built-in reconnect loop | Ensure WebSocket stability and run the submitter with automatic restarts (`systemd`/Kubernetes) |
### Enabling debug logs
Set the `LOG_LEVEL` environment variable to `debug` for verbose output:
```bash
LOG_LEVEL=debug bun tools/validator-set-submitter/main.ts run
```
Or in Docker/Kubernetes, add `LOG_LEVEL: "debug"` to the environment. Debug logs include per-tick skip reasons and detailed transaction information.