Backend support for #2151
Adds support for specifying custom behaviors via a list of strings.
When workflows are added or modified, minimal backend validation is done
to ensure that all custom behavior URLs are valid URLs (after removing
the git prefix and custom query arguments).
A separate `POST /crawlconfigs/validate/custom-behavior` endpoint is
also added, which can be used to validate a custom behavior URL. It
performs the same syntax check as above and then:
- For URL directly to behavior file, ensures URL resolves and returns a
2xx/3xx status code
- For Git repositories, uses `git ls-remote` to ensure they exist (and
that branch exists if specified)
---------
Co-authored-by: Ilya Kreymer <ikreymer@users.noreply.github.com>
* updates pydantic to 2.x
* also update to python 3.12
* additional type fixes:
- all Optional[] types must have a default value
- update to constrained types
- URL types converted from str
- test updates
Fixes#1940
Refactors backend deployment to:
- Use MOTOR_MAX_WORKERS (defaulting to 1) to reduce threads used by
mongodb connections
- Also sets backend workers to 1 by default to reduce default memory
usage
- Switches to gunicorn with uvloop worker for production use instead of
uvicorn (as recommended by uvicorn)
Lower thread count should address memory leak/increased usage, which
resulted in 5x thread x cpus x workers, eg. potentially 20 or 40 threads
just for mongodb connections. Lower default number of workers should
make it easier to scale backend with HPA if additional capacity.
Fixes#1467
* k8s local deployment work:
- make it easier to deploy w/o ingress by setting 'local_service_port' (suggested port 30870)
- if using local minio, ensure file endpoints set to /data/ and /data/ proxies correctly to local bucket
- if not using minio, ensure file endpoints point to correct access / endpoint url.
- setup should work with docker desktop, minikube, microk8s and k3s!
- nginx chart: bump nginx memory limit to 20Mi
- nginx image: 00-default-override-resolver-config -> 00-browsertrix-nginx-init for clarity
- nginx image: use default nginx.conf, pin to nginx 1.23.2
- mongo: readd readiness probe, bump connect wait timeout (needed for ci)
- config: set superadmin username to 'admin'
- config schema: set 'name' as required
- add sample chart values overrides:
- chart values: local-config.yaml for running locally with 'local_service_port'
- chart values: add microk8s-hosted.yaml for configuring a hosted microk8s setup
- chart values: add microk8s-ci.yaml for ci tests
- ci: remove docker swarm tests
- ci: add microk8s integration tests: launching cluster, logging in, running a crawl of example.com, downloading/checking WACZ
- bump to 1.1.0-beta.2
* simplify back to single config.env!
- back to good ole env vars!
- remove shared secret, which made it difficult to have scheduled crawls, since secrets are immutable, so could not update config if a scheduled crawl existed :/
- all env vars unified in configs/config.env - run-swarm.sh and run-pod.sh 'source' this config
- remove config.sample.yaml
- customize minio volume dir via config.env
- customize redis port via config.env
- include authsign ports in debug-ports config
* backend: refactor swarm support to also support podman (#260)
- implement podman support as subclass of swarm deployment
- podman is used when 'RUNTIME=podman' env var is set
- podman socket is mapped instead of docker socket
- podman-compose is used instead of docker-compose (though docker-compose works with podman, it does not support secrets, but podman-compose does)
- separate cli utils into SwarmRunner and PodmanRunner which extends it
- using config.yaml and config.env, both copied from sample versions
- work on simplifying config: add docker-compose.podman.yml and docker-compose.swarm.yml and signing and debug configs in ./configs
- add {build,run,stop}-{swarm,podman}.sh in scripts dir
- add init-configs, only copy if configs don't exist
- build local image use current version of podman, to support both podman 3.x and 4.x
- additional fixes for after testing podman on centos
- docs: update Deployment.md to cover swarm, podman, k8s deployment
- build backend and frontend with cacheing using GHA cache)
- streamline frontend image to reduce layers
- setup local swarm with test/setup.sh script, wait for containers to init
- copy sample config files as default (add storages.sample.yaml)
- add initial backend test for logging in with default superadmin credentials via 127.0.0.1:9871
- must use 127.0.0.1 instead of localhost for accessing frontend container within action
- use python-on-whale to use docker cli api directly, creating docker stack for each crawl or profile browser
- configure storages via storages.yaml secret
- add crawl_job, profile_job, splitting into base and k8s/swarm implementations
- split manager into base crawlmanager and k8s/swarm implementations
- swarm: load initial scale from db to avoid modifying fixed configs, in k8s, load from configmap
- swarm: support scheduled jobs via swarm-cronjob service
- remove docker dependencies (aiodocker, apscheduler, scheduling)
- swarm: when using local minio, expose via /data/ route in nginx via extra include (in k8s, include dir is empty and routing handled via ingress)
- k8s: cleanup minio chart: move init containers to minio.yaml
- swarm: stateful set implementation to be consistent with k8s scaling:
- don't use service replicas,
- create a unique service with '-N' appended and allocate unique volume for each replica
- allows crawl containers to be restarted w/o losing data
- add volume pruning background service, as volumes can be deleted only after service shuts down fully
- watch: fully simplify routing, route via replica index instead of ip for both k8s and swarm
- rename network btrix-cloud-net -> btrix-net to avoid conflict with compose network
- use statefulsets instead of deployments for mongo, redis, signer
- use k8s job + statefulset for running crawls
- use separate statefulset for crawl (scaled) and single-replica redis stateful set
- move crawl job update login to crawl_updater
- remove shared redis chart
package refactor:
- move to shared code to 'btrixcloud'
- move k8s to 'btrixcloud.k8s'
- move docker to 'btrixcloud.docker'
* apply /api prefix consistently, both directly through backend and when accessing via frontend, fixes#236
* docs: update local deployment docs to use 9871 instead of 8000, don't expose 8000 by default
* schemas: don't include /openapi.json as /healthz in documentation, keep /healthz at root
* k8s: route backend to /api without additional rewriting
- set WEB_CONCURRENCY env var to configure number of backend api workers for both docker and k8s
- set via 'backend_workers' in values.yaml
- also add 'rwp_base_url' to values.yaml
- update containers to use public webrecorder/browsertrix-backend and webrecorder/browsertrix-frontend containers
- make liveness, readiness and startup health checks more tolerant
* backend support for new watch system (#134):
- support for watch via redis pubsub and websocket connection to backend
- can support watch from any number of crawler instances to support scaled crawls
- use /archives/{aid}/crawls/{crawl_id}/watch/ws websocket endpoint
- ws: ignore graceful connectionclosedok exception, log other exceptions
- set logging to info to instead of debug for now (debug logs all ws traffic)
- remove old watch apis in backend
- remove old websocket routing to crawler instance for old watch system
- oauth bearer check: support websockets, use websocket object if no request object
- crawler args: replace --screencastPort with --screencastRedis
* support running backend + frontend together on k8s
* split nginx container into separate frontend service, which uses nignx-base image and the static frontend files
* add nginx-based frontend image to docker-compose build (for building only, docker-based combined deployment not yet supported)
* backend:
- fix paths for email templates
- chart: support '--set backend_only=1' and '--set frontend_only=1' to only force deploy one or the other
- run backend from root /api in uvicorn
