Releases: ipni/storetheindex
Releases · ipni/storetheindex
v0.4.11
What's Changed
Application
- Logging and config updates
- Update README.md
- Create config file description
- Fix version written by init
- Use default cidlink length by
- Update doc with current codes* Reframe API server
- Add logging around announce messages
- Cleanup policy evaluation
- Configure rate limiting and default values
Infrastructure and Deployment
- Deploy the latest release to
prod
environment - Point infra.cid.contact to
prod
running in sti AWS account - Scale up
dev
to 2 instances - Deploy latest to
dev
environment - Configure image update automation for
prod
- Fix regex replacement in automated PR creation template
- Deploy latest to
dev
environment - Increase memory and CPU for indexer on
prod
- Configure Promtail to forward host and container logs to PL loki
- Disable autoscaler in prod to investigate the selection policy
- Define
index-observer
ECR repo and CI to build and push images - Configure flux cd for
index-observer
ondev
cluster - Add
index-observer
to dev cluster manifests - Add encrypted credentials to access GitHub for
index-observer
- Stop deployment to
fil-dev
andfil-prod
via CircleCI - Onboard
autoretrieve
as a temporary tenant of stidev
cluster - dev-v2 on-boarding for gitops
- Allow GitHub Actions from ARG autoretrieve repo to publish to ECR
- Allow
put
via cloudfront sitting in front ofcid.contact
- Define dedicated KMS keys used to encrypt autoretrieve secrets
- Upgrade indexer config base layer
- Use larger instance types and override config in dev to limit providers
- Only whitelist ingestion from a specific provider in
dev
for testing - Fix invalid json in dev config
- Add memory optimized
r5a.2xlarge
workers todev
Full Changelog: v0.4.10...v0.4.11
v0.4.10
What's Changed
- Fix goroutine accumulation when announces arrive how peer that is busy syncing
- Handle direct announce requests the same as those received over pubsub
- Support HTTP addresses in pubsub announce messages
- Support multiple addresses in announce messages
Full Changelog: v0.4.9...v0.4.10
v0.4.9
What's Changed
- If no value for "indexer" CLI flag, then read from config
- Update storethehash to get index recovery fix
Infrastructure changes
- Send metrics to AWS managed prometheus via private link
- Define base promtail kustomization used for log forwarding
- Publish containers to ECR repo when a release is created
- Spread indexer instances across AZs when possible
- Create pl-grafana IAM user to integrate with PL Grafana
- Write
prod
metrics to a dedicated AWS managed prometheus - Configure resources for
dev
amdprod
Full Changelog: v0.4.8...v0.4.9
v0.4.8
What's Changed
- Use new flag names in commands
- Update to core with storehash panic fix
Full Changelog: v0.4.7...v0.4.8
v0.4.7
What's Changed
- New storethehash valuestore with automatic index garbage collection.
- Automatic upgrade of storethehash index to new format
- Fix wrong publisher address when using HTTP
- Support a list of peers from which we accept advertisements for other providers
- Upgrade config using
./storetheindex init --upgrade
- Revised command flags to differentiate publisher and provider peer id flags
Infrastructure and Deployment Changes
- Define prod kustomization overlay
- Define production infrastructure in
us-east-2
- Configure EBS CSI controller to run on all nodes
- Bump github.com/ipfs/go-ipfs from 0.11.0 to 0.12.1
- Set up public ECR and publish containers via CI
- Remove redundant private ECR repo
- Revert "Remove redundant private ECR repo"
- Remove public ECR and publish images to the private repo instead
- Automate deployment on container image publication in
dev
- Do not publish to ECR if pusher is
sti-bot
- Configure readiness probe and pod disruption budget
- Remove skip message from image update automation
- Fix ECR build still running on sti bot pushes to
deploy
directory - Use io2 storage class for indexer pods in
dev
- Add AWS EBS CSI to dev cluster manifests
- Use
io2
storage class forprod
instances
Full Changelog: v0.4.6...v0.4.7
v0.4.3
- fix panic in registry: #320
What's Changed
- Give Kyle access to
dev
cluster by @masih in #309 - Set up caching layer for dev endpoint using CloudFront by @masih in #301
- Added cmharden and steveFraser to access K8s by @cmharden in #313
- Cache 404s in cloudfront by @MarcoPolo in #314
- Set up cluster-autoscaler in
dev
by @masih in #308 - Add
cluster-autoscaler
to list of cluster level kustomizations by @masih in #316 - Fix out of bounds panic in registery by @gammazero in #320
- sync: update CI config files by @web3-bot in #317
- Allow Kyle to interact with dev KMS by @masih in #326
- Set up K8S monitoring for
dev
cluster by @masih in #325 - Fix
MalformedYAMLError
when kustomization uses HTTPS to clone repo by @masih in #329
Full Changelog: v0.4.1...v0.4.3
v0.4.2
v0.4.1
What's Changed
- Fix find latency metric
- Store publisher address with publisher ID
- Update dependencies on go-legs, go-datatransfer, go-graphsync
- Fix pprof handlers
Full Changelog: v0.4.0...v0.4.1
v0.4.0
What's Changed
- Use IPLD bindnode and reduce boilerplate code in ingest schema
- Fix goroutine leak in go-statemachine
- More efficient batch processing of multihashes
Full Changelog: v0.3.8...v0.4.0
v0.3.8
- Limit number of open connection to the http serve
- Fix goroutine leak in go-libp2p-swarm
- Remove recursion depth limits for sync
- Fair sharing of ingest workers to publishers
- Drop support for Go 1.16
- Set the maximum acceptable length for advertisement metadata
- Fix issue with concurrent syncs to same publisher: don't optimistically delete ad
- If failure during entry sync, do not record all ads a being synced
- Docker build uses go 1.18