- By default, use only `ingressClassName` for ingress class name and
corresponding field in cert-manager
- Only use old 'kubernetes.io/ingress.class' if
ingress.useOldClassAnnotation is set
- Allow for using old annotation only for backwards compatibility, eg.
for GCP
- Closes#2267 and #1570
---------
Co-authored-by: Ilya Kreymer <ikreymer@gmail.com>
- Renames `inject_analytics` to `inject_extra` and updates docs
- Manually tracks page views to enable passing custom props
- Tracks copying collection share link and downloading a public
collection
---------
Co-authored-by: emma <hi@emma.cafe>
Fixes#2182
This rather large PR adds the rest of what should be needed for public
collections work in the frontend.
New API endpoints include:
- Public collections endpoints: GET, streaming download
- Paginated list of URLs in collection with snapshot (page) info for
each
- Collection endpoint to set home URL
- Collection endpoint to upload thumbnail as stream
- DELETE endpoint to remove collection thumbnail
Changes to existing API endpoints include:
- Paginating public collection list results
- Several `pages` endpoints that previously only supported `/crawls/` in
their path, e.g. `/orgs/{oid}/crawls/all/pages/reAdd`, now support
`/uploads/` and `/all-crawls/` namespaces as well. This is necessitated
by adding pages for uploads to the database (see below). For
`/orgs/{oid}/namespace/all/pages/reAdd`, `crawls` or `uploads` will
serve as a filter to only affect crawls of that given type. Other
endpoints are more liberal at this point, and will perform the same
action regardless of the namespace used in the route (we'll likely want
to change this in a follow-up to be more consistent).
- `/orgs/{oid}/namespace/all/pages/reAdd` now kicks off a background job
rather than doing all of the computation in an asyncio task in the
backend container. The background job additionally updates collection
date ranges, page/size counts, and tags for each collection in the org
after pages have been (re)added.
Other big changes:
- New uploads will now have their pages read into the database!
Collection page counts now also include uploads
- A migration was added to start a background job for each org that will
add the pages for previously-uploaded WACZ files to the database and
update collections accordingly
- Adds a new `ImageFile` subclass of `BaseFile` for thumbnails that we
can use for other user-uploaded image files moving forward, with
separate output models for authenticated and public endpoints
Fixes#2170
The number of days to delay file replication deletion by is configurable
in the Helm chart with `replica_deletion_delay_days` (set by default to
7 days in `values.yaml` to encourage good practice, though we could
change this).
When `replica_deletion_delay_days` is set to an int above 0, when a
delete replica job would otherwise be started as a Kubernetes Job,
a CronJob is created instead with a cron schedule set to run yearly,
starting x days from the current moment. This cronjob is then deleted by
the operator after the job successfully completes. If a failed
background job is retried, it is re-run immediately as a Job rather
than being scheduled out into the future again.
---------
Co-authored-by: Ilya Kreymer <ikreymer@users.noreply.github.com>
- By default, all locales are enabled to make it easy for local deployments to test new locales
- Adds DE, FR, PT locales to make way for translation in Weblate
Closes#2223
- [x] Adds `localesAvailable` to `/api/settings` endpoint, and uses that
list if available, rather than the full list of translated locales, to
determine which options to display to users
- [x] ~~Uses the user's browser locales, filtered to the current
language setting, for formatting numbers, dates, and durations~~
- [x] Adds & persists checkbox for "use same language for formatting
dates and numbers" in user settings
- [x] Replaces uses of `sl-format-bytes` with `localize.bytes(...)`, and
`sl-format-date` with replacement `btrix-format-date` that properly
handles fallback locales
- [x] Caches all number/duration/datetime formatters by a combined key
consisting of app language, browser language, browser setting, and
formatter options so that all formatters can be reused if needed
(previously any formatter with non-default options would be recreated
every render)
- [x] Splits out ordinal formatting from number formatter, as it didn't
make much sense in some non-English locales
- [x] Adds a little demo of date/time/duration/number formatting so you
can see what effect your language settings have
https://github.com/user-attachments/assets/724858cb-b140-4d72-a38d-83f602c71bc7
---------
Signed-off-by: emma <hi@emma.cafe>
Co-authored-by: Ilya Kreymer <ikreymer@gmail.com>
Co-authored-by: Ilya Kreymer <ikreymer@users.noreply.github.com>
Adds sending a cancellation email when a subscription is cancelled.
- The email may also include an option survey optional survey URL, if
configured in helm chart `survey_url` setting.
- Cancellation e-mail configured in `sub_cancel` e-mail template
- E-mails are sent to all org admins.
- Also adds `trialing_canceled` subscription state to differentiate from
a default `trialing` which will automatically rollover into `active`.
- The email is sent when: a new cancellation date is added for an
`active` subscription, or a `trialing` subscription is changed to to
`trialing_canceled`. (A subscription can be canceled/uncanceled several
times before actual date, and e-mail is sent every time it is canceled.)
- The 'You have X days left of your trial' is also always displayed when
state is in trialing_canceled.
Fixes#2229
---------
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
Closes#2222
Adds a runtime script that gets set to either inject the plausible
script tags, or do nothing, that runs at initialization of the frontend
container.
Fixes#2186
Background job emails will no longer fail to send for jobs unrelated to
file replication or replica deletion.
Also uses `AnyJob` for paginated background job response model, to fix
typing being out of data following addition of other types of background
jobs and lower overhead for adding new ones moving forward.
Fixes#2106
Docs are now hosted as part of the frontend at `/docs` by default.
- If `docs_url` is set in the helm chart, the `/docs` endpoint will
redirect to that endpoint instead
- Use multi-stage python image to build mkdocs as part of frontend, then
copy static output
- Dir layout: mkdocs.yml and docs into frontend/docs
- CI: Update docs build GH action to use new path
- Update all frontend paths to use `/docs/` instead of
`https://docs.browsertrix.com/`
---------
Co-authored-by: Henry Wilkinson <henry@wilkinson.graphics>
Resolves#1354
Supports crawling through pre-configured proxy servers, allowing users to select which proxy servers to use (requires browsertrix crawler 1.3+)
Config:
- proxies defined in btrix-proxies subchart
- can be configured via btrix-proxies key or separate proxies.yaml file via separate subchart
- proxies list refreshed automatically if crawler_proxies.json changes if subchart is deployed
- support for ssh and socks5 proxies
- proxy keys added to secrets in subchart
- support for default proxy to be always used if no other proxy configured, prevent starting cluster if default proxy not available
- prevent starting manual crawl if previously configured proxy is no longer available, return error
- force 'btrix' username and group name on browsertrix-crawler non-root user to support ssh
Operator:
- support crawling through proxies, pass proxyId in CrawlJob
- support running profile browsers which designated proxy, pass proxyId to ProfileJob
- prevent starting scheduled crawl if previously configured proxy is no longer available
API / Access:
- /api/orgs/all/crawlconfigs/crawler-proxies - get all proxies (superadmin only)
- /api/orgs/{oid}/crawlconfigs/crawler-proxies - get proxies available to particular org
- /api/orgs/{oid}/proxies - update allowed proxies for particular org (superadmin only)
- superadmin can configure which orgs can use which proxies, stored on the org
- superadmin can also allow an org to access all 'shared' proxies, to avoid having to allow a shared proxy on each org.
UI:
- Superadmin has 'Edit Proxies' dialog to configure for each org if it has: dedicated proxies, has access to shared proxies.
- User can select a proxy in Crawl Workflow browser settings
- Users can choose to launch a browser profile with a particular proxy
- Display which proxy is used to create profile in profile selector
- Users can choose with default proxy to use for new workflows in Crawling Defaults
---------
Co-authored-by: Ilya Kreymer <ikreymer@gmail.com>
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
Attempt to auto-adjust PVC storage if:
- used storage (as reported in redis by the crawler) * 2.5 >
total_storage
- will cause PVC to resize, if possible (not supported by all drivers)
- uses multiples of 1Gi, rounding up to next GB
- AVAIL_STORAGE_RATIO hard-coded to 2.5 for now, to account for 2x space
for WACZ plus change for fast updating crawls
Some caveats:
- only works if the storageClass used for PVCs has
`allowVolumeExpansion: true`, if not, it will have no effect
- designed as a last resort option: the `crawl_storage` in values and
`--sizeLimit` and `--diskUtilization` should generally result in this
not being needed.
- can be useful in cases where a crawl is rapidly capturing a lot of
content in one page, and there's no time to interrupt / restart, since
the other limits apply only at page end.
- May want to have crawler update the disk usage more frequently, not
just at page end to make this more effective.
Tweaks to how execution time is tracked for more accuracy + excluding
waiting states:
- don't update if crawl state is in a 'waiting state' (waiting for
capacity or waiting for org limit)
- rename start states -> waiting states for clarity
- reset lastUpdatedTime if two consecutive updates of non-running state,
to ensure non-running states don't count, but also account for
occasional hiccups -- if only one update detects non-running state,
don't reset
- webhooks: move start webhook to when crawl actually starts for first
time (db lastUpdatedTime is not yet + crawl is running)
- don't set lastUpdatedTime until pods actually running
- set crawljob update interval to every 10 seconds for more accurate
execution time tracking
- frontend: show seconds in 'Execution Time' display
If a cronjob is disabled, the operator should quickly return a success
value so that the job can be terminated.
Was previously returning an incorrect response, causing disabled
cronjobs to not be cleaned up. Add proper typing to always return correct response
- make crawl args a reusable template
- adds QA_ARGS to configmap, setting to same value as CRAWL_ARGS but
with --behaviors= prepended to disable behaviors for QA, to improve
performance of QA runs.
fixes#1962
- only enable if 'enable_auto_resize' is true, default to false
- if true, set memory limit to 1.2 of memory requests, resize when
hitting 'soft oom' of initial request, adjust by 1.2 (current behavior)
up to max_crawler_memory
- if false, set memory limit to max_crawler_memory and never adjust
memory requests or memory limits
- part of #1959
Fixes https://github.com/webrecorder/browsertrix/issues/1905
- adds a new top-level `/api/subscriptions` endpoint and SubOps handler on
the backend.
- enable subscriptions API endpoints available only if `billing_enabled` is
set in helm chart
- new POST /subscriptions/create, /subscriptions/update,
/subscriptions/cancel API endpoints
- Subscriptions mongo collection storing timestamped /subscription
API events
- GET /subscriptions/events API to get subscription events, support for filtering and sorting
- Subscription data model
- Support for setting and handling readOnlyOnCancel on org
- /orgs/<id>/billing-portal to lookup portalUrl using external API
- subscription in org getter and list views
- mark org as readOnly for subscription status `paused_payment_failed`, clears it on status `active`
---------
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
Limit egress traffic from crawler/profilebrowser pods to the internet
and limited internal services like dns, redis, frontend, auth-signer on certain ports
---------
Co-authored-by: Ilya Kreymer <ikreymer@gmail.com>
Fixes#1432
Refactors the invite + registration system to be simpler and more consistent
with regards to existing user invites. Previously, per-user invites are
stored in the user.invites dict instead of in the invites collection,
which creates a few issues:
- Existing user do not show up in Org Invites list: #1432
- Existing user invites also do not expire, unlike new user invites,
creating potential security issue.
Instead, existing user invites should be treated like new user invites.
This PR moves them into the same collection,
adding a `userid` field to InvitePending to match with an existing user.
If a user already exists, it will be matched by userid, instead of by
email. This allows for user to update their email while still being
invited. Note that the email of the invited existing user will not
change in the invite email. This is also by design: an admin of one org
should not be given any hint that an invited user already has an
account, such as by having their email automatically update. For an org
admin, the invite to a new or existing user should be indistinguishable.
The sha256 of invite token is stored instead of actual token for better
security.
The registration system has also been refactored with the following
changes:
- Auto-creation of new orgs for new users has been removed
- User.create_user() replaces the old User._create() and just creates the user with
additional complex logic around org auto-add
- Users are added to org in org add_user_to_org()
- Users are added to org through invites with add_user_with_invite()
Tests:
- Additional tests include verifying that existing and new pending
invites appear in the pending invites list
- Tests for `/users/invite/<token>?email=` and
`/users/me/invite/<token>` endpoints
- Deleting pending invites
- Additional tests added for user self-registration, including existing
user self-registration to default org of existing user (in nightly
tests)
Fixes#890
This PR introduces new streaming superuser-only API endpoints to export
and import database information for an organization. New Adminstrator
deployment documentation on how to manage the process and copy files
between S3 buckets as needed is also included.
---------
Co-authored-by: Henry Wilkinson <henry@wilkinson.graphics>
Co-authored-by: Ilya Kreymer <ikreymer@gmail.com>
Updates the /api/orgs/create endpoint to:
- not have name / slug be required, will be renamed on first user via
#1870
- support optional quotas
- support optional first admin user email, who will receive an invite to
join the org.
Also supports a new shared secret mechanism, to allow an external
automation to access the /api/orgs/create endpoint (and only that
endpoint thus far) via a shared secret instead of normal login.
Fixes#1893
- Removes crawl workflow-scoped configmaps, and replaces with operator-controlled
per-crawl configmaps that only contain the json config passed to Browsertrix
Crawler (as a volume).
- Other configmap settings replaced are replaced the custom CrawlJob options
(mostly already were, just added profile_filename and storage_filename)
- Cron jobs also updated to create CrawlJob without relying on configmaps,
querying the db for additional settings.
- The `userid` associated with cron jobs is set to the user that last modified
the schedule of the crawl, rather than whomever last modified the workflow
- Various functions that deal with updating configmaps have been removed,
including in migrations.
- New migration 0029 added to remove all crawl workflow configmaps
- needed to support js-wacz signing requests in upcoming crawler versions
- Also has slightly increased memory requirements due to new versions of
some libraries.
- 0.5.2 adds a fix to dropping the fractional part of the second, to make
it work with ISO date strings that have microseconds, such as those from
js-wacz.
Resolves https://github.com/webrecorder/browsertrix/issues/1874
Support for new two-part sign up flow if first admin user is added to org
- If new user, user registers first, then is able to change the org name / slug on following screen
- If existing user, user accepts invite, then is able to change the org name / slug on following screen
- After confirming org slug name, user is taken to dashboard, or error is shown if org name or slug already taken.
- If org name == org id, org name and slug is automatically set to `{Your Name}'s Archive` when first user is registered / accepts invite
- Email templates updated to better reflect new / existing users and not show org name if it is 'unset' (org name == org id internally)
- tests: frontend unit testing for accept + invite screens.
---------
Co-authored-by: Ilya Kreymer <ikreymer@gmail.com>
Co-authored-by: sua yoo <sua@suayoo.com>
Co-authored-by: sua yoo <sua@webrecorder.org>
Co-authored-by: Henry Wilkinson <henry@wilkinson.graphics>
Co-authored-by: Ilya Kreymer <ikreymer@users.noreply.github.com>
Co-authored-by: Emma Segal-Grossman <hi@emma.cafe>
Fixes#1888
Refactors scale handling:
- Ensures number of scaled instances does not exceed number of pages,
but is also at minimum 1
- Checks for finish condition to be numFailed + numDone >= desired scale
- If at least one instance succeeds, crawl considers successful / done.
- If all instances fail, crawl considered failed
- Ensures that pod done count >= redis done count
---------
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
- Adds Browsertrix GitHub repo and Webrecorder forum to the bottom of
the support email.
- Adds note about having an applicable plan to contact support
---------
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
Backend work for first two tasks of
https://github.com/webrecorder/browsertrix/issues/1875
New /billing API endpoint to be added separately once we have a better
idea of what data we can get from the payment processor.
- add a 'expire_at_duration_seconds' which is 75% of actual presign
duration time, or <25% remaining until presigned URL actually expires to
ensure presigned URLs are updated early than when they actually expire
- set cached expireAt time to the renew at time for more frequent
updates
- update QA configmap in place with updated presigned URLs when expireAt
time is reached
- mount qa config volume under /tmp/qa/ without subPath to get automatic
updates, which crawler will handle
- tests: fix qa test typo (from main)
- fixes#1864
- allow configuring QA run scale via 'qa_scale' setting in helm values
(overriding any setting on the qa crawljob)
- adds additional comments to browser instances helm values settings for clarity
- fixes#1842
Currently, the workflow crawl settings were not being included at all in
QA runs.
This mounts the crawl workflow config, as well as QA configmap, into QA
run crawls, allowing for page limits from crawl workflow to be applied
to QA runs.
It also allows a different number of browser instances to be used for QA
runs, as QA runs might work better with less browsers, (eg. 2 instead of
4). This can be set with `qa_browser_instances` in helm chart.
Default qa browser workers to 1 if unset (for now, for best results)
Fixes#1828
- ensure max_crawler_memory_size is inited before it is set!
- pass profile_browser_memory / profile_browser_cpu from chart values
- map volume to /tmp/home to avoid persisting /tmp for profiles
Adds a `max_crawler_memory` chart setting, which, if set, will
defines the upper crawler memory limit that crawler pods can be resized up to.
If not set, auto resizing is disabled and pods are always set to 'crawler_memory' memory
if 'registration_enabled' is set, check 'registration_org_id' for org id
of an existing org that new users should be added to when they register.
if omitted, default to the default org
Fixes#1729
Repository Index: Generate an index.yaml in ./docx/helm-repo/index.yaml
to allow for browsertrix to be a helm repository.
docs: rename docs.browsertrix.cloud -> docs.browsertrix.com
docs: update deployment doc to mention helm repo as preferred way to
install
docs build action: generate repository index in GH action
publish action: update auto-generated message to mention installing from
the repo.
---------
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
- fixes#1684
- can be used to optionally restrict QA to only some crawls (eg. with
browsertrix-crawler>=1.0.0)
- enforce error on backend (return 400) and handle special error on the
frontend
- Remove globals from profile, uploads, and qa test modules in favor of fixtures
- Add retries to fix intermittent test failures due to timing
---------
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
Co-authored-by: Ilya Kreymer <ikreymer@gmail.com>
- priority classes <-10 are ignored by cluster-autoscaler so QA jobs
with too low priorities never run
- start crawl priorities at 0 going down (same as before)
- start qa run priorities at -2 going down (instead of -100)
- this means a crawl of with scale of 3 can be preempted by 1st qa pod,
but otherwise crawls have higher priority
- rename priority classes as they are otherwise immutable and error on
helm upgrade
This allows for more room in lower pri classes for other type of
objects, while keeping in mind the -10 and below threshold: (see:
https://github.com/kubernetes/autoscaler/blob/master/cluster-autoscaler/FAQ.md)
- modify invite email template to answer common questions
- email templates: make each email template overridable with --set-file
- docs: update customization doc to document how to customize email
templates
---------
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
Supports horizontal pod autoscaling (hpa) for backend and frontend pods:
- use cpu and memory averages
- adjust base memory + cpu for backend
- threshold set to 80% cpu and 95% memory utilization by default
(configurable in values.yaml)
- instead of backend and frontend replicas, set max replicas in
values.yaml
- only enable hpa if backend_max_replicas or frontend_max_replicas is
>1, default to 1 for now
- set memory limit to 1.2x memory request to provide extra padding and
avoid OOM
- attempt to resize crawler pods by 1.2x when exceeding 90% of available
memory
- do a 'soft OOM' (send extra SIGTERM) to pod when reaching 100% of
requested memory, resulting in faster graceful restart, but avoiding a
system-instant OOM Kill
- Fixes#1632
---------
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
Supports running QA Runs via the QA API!
Builds on top of the `issue-1498-crawl-qa-backend-support` branch, fixes
#1498
Also requires the latest Browsertrix Crawler 1.1.0+ (from
webrecorder/browsertrix-crawler#469 branch)
Notable changes:
- QARun objects contain info about QA runs, which are crawls
performed on data loaded from existing crawls.
- Various crawl db operations can be performed on either the crawl or
`qa.` object, and core crawl fields have been moved to CoreCrawlable.
- While running,`QARun` data stored in a single `qa` object, while
finished qa runs are added to `qaFinished` dictionary on the Crawl. The
QA list API returns data from the finished list, sorted by most recent
first.
- Includes additional type fixes / type safety, especially around
BaseCrawl / Crawl / UploadedCrawl functionality, also creating specific
get_upload(), get_basecrawl(), get_crawl() getters for internal use and
get_crawl_out() for API
- Support filtering and sorting pages via `qaFilterBy` (screenshotMatch, textMatch)
along with `gt`, `lt`, `gte`, `lte` params to return pages based on QA results.
---------
Co-authored-by: Tessa Walsh <tessa@bitarchivist.net>
Fixes#1597
New endpoints (replacing old migration) to re-add crawl pages to db from
WACZs.
After a few implementation attempts, we settled on using
[remotezip](https://github.com/gtsystem/python-remotezip) to handle
parsing of the zip files and streaming their contents line-by-line for
pages. I've also modified the sync log streaming to use remotezip as
well, which allows us to remove our own zip module and let remotezip
handle the complexity of parsing zip files.
Database inserts for pages from WACZs are batched 100 at a time to help
speed up the endpoint, and the task is kicked off using
asyncio.create_task so as not to block before giving a response.
StorageOps now contains a method for streaming the bytes of any file in
a remote WACZ, requiring only the presigned URL for the WACZ and the
name of the file to stream.
- Backend: Use separate resource constraints for profiles: default
profile browser resources to either 'profile_browser_cpu' /
'profile_browser_memory' or single browser 'crawler_memory_base' /
'crawler_cpu_base', instead of scaled to the number of browser workers
- Frontend: check that profile html page is loading, keep retrying if
still getting nginx error instead of loading an iframe with the error.
Fixes#1598 (Copy of #1599 from 1.9.4)
Part of #1241
### Changes
- Renames all instances of "Browsertrix Cloud" to "Browsertrix" on the
front end, emails, and documentation
---------
Co-authored-by: emma <hi@emma.cafe>
- increases the failureThreshold for startupProbe for the api backend
container to account for long running migrations, upto 300 seconds
- add `/healthzStartup` which checks if db is ready
- bump
- keeps `/healthz` to always return 200 when running
- increases livenessProbe failureThreshold to be higher than readiness
probe, following recommended best practice of liveness probe > readiness
probe
- fixes#1559
Fixes#1502
- Adds pages to database as they get added to Redis during crawl
- Adds migration to add pages to database for older crawls from
pages.jsonl and extraPages.jsonl files in WACZ
- Adds GET, list GET, and PATCH update endpoints for pages
- Adds POST (add), PATCH, and POST (delete) endpoints for page notes,
each with their own id, timestamp, and user info in addition to text
- Adds page_ops methods for 1. adding resources/urls to page, and 2.
adding automated heuristics and supplemental info (mime, type, etc.) to
page (for use in crawl QA job)
- Modifies `Migration` class to accept kwargs so that we can pass in ops
classes as needed for migrations
- Deletes WACZ files and pages from database for failed crawls during
crawl_finished process
- Deletes crawl pages when a crawl is deleted
Note: Requires a crawler version 1.0.0 beta3 or later, with support for
`--writePagesToRedis` to populate pages at crawl completion. Beta 4 is
configured in the test chart, which should be upgraded to stable 1.0.0
when it's released.
Connected to https://github.com/webrecorder/browsertrix-crawler/pull/464
---------
Co-authored-by: Ilya Kreymer <ikreymer@gmail.com>
Supports setting WARC prefix for WARCs inside WACZ to `<org slug>-<slug
[crawl name | first seed host]>`.
- Prefix set via WARC_PREFIX env var, supported in browsertrix-crawler
1.0.0-beta.4 or higher
If crawl name is provided, uses crawl name, other hostname of first
seed. The name is 'sluggified', using lowercase alphanum characters
separated by dashes.
Ex: in an organization called `Default Org`, a crawl of
`https://specs.webrecorder.net/` and no name will have WARCs named:
`default-org-specs-webrecorder-net-....warc.gz`
If the crawl is given the name `SPECS`, the WARCs will be named
`default-org-specs-manual-....warc.gz`
Fixes#412 in a default way.
configmap: add --screenshot thumbnail,view as default screenshots
version: update update-version.sh to add newline in version.py to match
new black formatting (from changes in #1507)
Fixes#1519
Fixes#1341
Adds "User Agent" field to workflow editor under the Browser Settings
tab. If not set, the crawler will use the browser's default user agent.
Also added to docs and to the workflow details page (if set).
---------
Co-authored-by: Henry Wilkinson <henry@wilkinson.graphics>
Co-authored-by: Ilya Kreymer <ikreymer@users.noreply.github.com>
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
Fixes#1385
## Changes
Supports multiple crawler 'channels' which can be configured to
different browsertrix-crawler versions
- Replaces `crawler_image` in helm chart with `crawler_channels` array
similar to how storages are handled
- The `default` crawler channel must always be provided and specifies
the default crawler image
- Adds backend `/orgs/{oid}/crawlconfigs/crawler-channels` API endpoint
to fetch information about available crawler versions (name, image, and
label) and test
- Adds crawler channel select to workflow creation/edit screens and
profile creation dialog, and updates related API endpoints and
configmaps accordingly. The select dropdown is shown only if more than
one channel is configured.
- Adds `crawlerChannel` to workflow and crawl details.
- Add `image` to crawler image, used to display actual image used as
part of the crawl.
- Modifies `crawler_crawl_id` backend test fixture to use `test` crawler
version to ensure crawler versions other than latest work
- Adds migration to add `crawlerChannel` set to `default` to existing
workflow and profile objects and workflow configmaps
---------
Co-authored-by: Ilya Kreymer <ikreymer@gmail.com>
Co-authored-by: Henry Wilkinson <henry@wilkinson.graphics>
