Skip to content

Commit

Permalink
Merge branch 'master' of github.com:Yelp/paasta into u/krall/PAASTA-1…
Browse files Browse the repository at this point in the history
…8298_delete_forecast_policy_decision_policy
  • Loading branch information
EvanKrall committed Dec 3, 2024
2 parents f0277b5 + ef0195f commit 998333a
Show file tree
Hide file tree
Showing 57 changed files with 1,361 additions and 966 deletions.
99 changes: 99 additions & 0 deletions CODEOWNERS
Validating CODEOWNERS rules …
Original file line number Diff line number Diff line change
@@ -0,0 +1,99 @@
# NOTE: "we" in this file will refer to the Compute Infrastructure team at Yelp

# prevent cheeky modifications :)
CODEOWNERS @Yelp/paasta

## These impact the build process, so we probably want CI review on 'em
**/Makefile @Yelp/paasta
Makefile @Yelp/paasta
# this one is kinda misnamed since it's really a patch/ directory :p
contrib/ @Yelp/paasta
tox.ini @Yelp/paasta
# NOTE: we should probably CODEOWNERS requirements, but this would slow down ML
# Compute since they have a lot of logic inside service_configuration_lib and
# tend to make PaaSTA releases that just bump this library
# XXX: maybe CODEOWNERS to CI + ML Compute?

## These contain entrypoints for PaaSTA or otherwise control entrypoint builds
setup.py @Yelp/paasta
debian/control @Yelp/paasta
debian/paasta-tools.links @Yelp/paasta
debian/paasta-tools.substvars @Yelp/paasta
debian/rules @Yelp/paasta
# this one is unfortunate since operators own the paasta status code for their
# workloads, but atm all this is in one large file
paasta_tools/cli/ @Yelp/paasta
# exclude these files from being codeownered to us
paasta_tools/cli/cmds/spark_run.py
paasta_tools/cli/cmds/security_check.py


# we don't own all the API code, but we do own most of it. if this turns into a
# problem, we can split it up further or see if we can specifically remove
# CODEOWNERS from the parts we don't own
paasta_tools/api/ @Yelp/paasta
paasta_tools/api/* @Yelp/paasta

## common code used in many operators OR core code for paasta services
# at the time or writing, we only care/use pause_service_autoscaler.py and
# utils.py, but we might as well own the whole directory
paasta_tools/autoscaling/ @Yelp/paasta
# this might not always be true, but for now any file that starts with setup
# is pretty foundational (and do things like create k8s deployments, CRDs,
# etc.)
paasta_tools/setup* @Yelp/paasta
# and same logic for files that start with cleanup
paasta_tools/cleanup* @Yelp/paasta
# ...too bad we didn't call this cleanup_completed_pods.py :p
paasta_tools/prune_completed_pods.py @Yelp/paasta
# atm, the only thing we care about/use here is the kubernetes.py file, but
# just in case we refactor things out a bit
paasta_tools/instance/ @Yelp/paasta
# helpers or base classes used in many places
paasta_tools/*utils.py @Yelp/paasta
paasta_tools/kubernetes/ @Yelp/paasta
paasta_tools/long_running_service_tools.py @Yelp/paasta
# metadata generators - for now all of these are important enough that we
# should be involved
paasta_tools/generate* @Yelp/paasta
# PaaSTA/Tron core code
paasta_tools/adhoc_tools.py @Yelp/paasta
paasta_tools/kubernetes_tools.py @Yelp/paasta
paasta_tools/eks_tools.py @Yelp/paasta
paasta_tools/paasta* @Yelp/paasta
# mesh-related code
paasta_tools/envoy_tools.py@Yelp/paasta
paasta_tools/hacheck.py @Yelp/paasta
paasta_tools/smartstack_tools.py @Yelp/paasta
# same logic as other wildcards above - currently all of these are important
# and owner by us
paasta_tools/list* @Yelp/paasta
paasta_tools/tron/ @Yelp/paasta
paasta_tools/tron* @Yelp/paasta
# secret management
paasta_tools/secret_providers/ @Yelp/paasta
paasta_tools/secret_tools.py @Yelp/paasta

## directories we want to break apart
# we have a bad habit of depending on things in here - which is always a fun
# time since we're pretty lax with the code quality of things we add here
# initially. hopefully adding CODEOWNERS will help us be more mindful and
# eventually move these to more appropriate/tested locations (as well as
# protect important files like the autotune or SSR management code)
paasta_tools/contrib/* @Yelp/paasta

## files used for alerting
# NOTE: we could codeowner check*, but there's enough scripts like
# check_flink_services_health.py, check_spark_jobs.py, etc. that we
# should break this down to just what we own
paasta_tools/check_autoscaler_max_instances.py @Yelp/paasta
paasta_tools/check_kubernetes_api.py @Yelp/paasta
paasta_tools/check_kubernetes_services_replication.py @Yelp/paasta
paasta_tools/check_oom_events.py @Yelp/paasta
paasta_tools/check_services_replication_tools.py @Yelp/paasta
paasta_tools/monitoring/ @Yelp/paasta
paasta_tools/monitoring* @Yelp/paasta
paasta_tools/oom_logger.py @Yelp/paasta
# not technically alerting, but very useful for spot interruption handling
# so we should be careful with it
paasta_tools/broadcast_log_to_services.py @Yelp/paasta
181 changes: 181 additions & 0 deletions debian/changelog
Original file line number Diff line number Diff line change
@@ -1,3 +1,184 @@
paasta-tools (1.8.2) xenial; urgency=medium

* 1.8.2 tagged with 'make release'
Commit: Merge pull request #3989 from
Yelp/u/jfong/fix_paasta_status_bouncestatus Fix backwards "Bouncing
to" status

-- Jen Patague <[email protected]> Mon, 02 Dec 2024 11:59:05 -0800

paasta-tools (1.8.1) xenial; urgency=medium

* 1.8.1 tagged with 'make release'
Commit: Enable topology spread constraints by default for tron
(#3987) I debated getting rid of this toggle entirely, but I guess
it"s nice to keep around just in case? I"ll also follow this up
with a Puppet PR to clean things up there

-- Luis Perez <[email protected]> Thu, 14 Nov 2024 12:00:34 -0800

paasta-tools (1.8.0) xenial; urgency=medium

* 1.8.0 tagged with 'make release'
Commit: Add tron topology_spread_constraints support to PaaSTA
(#3983) This adds support for configuring Tron-launched pods with a
default Topology Spread Constraint (and node affinities) that will
spread pods out across multiple AZs - otherwise, Karpenter will
overwhelmingly favor a single AZ due to our config

-- Luis Perez <[email protected]> Mon, 11 Nov 2024 12:36:11 -0800

paasta-tools (1.7.2) xenial; urgency=medium

* 1.7.2 tagged with 'make release'
Commit: Allowlist eks files in config_utils (#3984)
updater.write_configs() is not just used for managing autotune files
- SSR shard creation uses it to manage creating/updating the non-
autotuned_defaults config files (which are now always eks-
CLUSTER.yaml).

-- Luis Perez <[email protected]> Tue, 29 Oct 2024 13:52:41 -0700

paasta-tools (1.7.1) xenial; urgency=medium

* 1.7.1 tagged with 'make release'
Commit: Bump service-configuration-lib to v3 to stop mounting
/etc/pki/spark (#3982) This is no longer needed since we"ve been
running Spark on just EKS for a while.

-- Luis Perez <[email protected]> Wed, 16 Oct 2024 13:53:57 -0700

paasta-tools (1.7.0) xenial; urgency=medium

* 1.7.0 tagged with 'make release'
Commit: Cleaning up Mesos from paasta readthedocs - PAASTA-18313
(#3954) * Cleaning up Mesos from paasta readthedocs * Address
reviews * Address more reviews * Addressing yelpsoa files reviews

-- Eman Elsabban <[email protected]> Wed, 16 Oct 2024 11:32:36 -0700

paasta-tools (1.6.4) xenial; urgency=medium

* 1.6.4 tagged with 'make release'
Commit: Merge pull request #3977 from
Yelp/u/calvinli/automated_redeploys_default_toggle let
enable_automated_redeploys default value be configurable

-- Evan Krall <[email protected]> Wed, 09 Oct 2024 15:15:13 -0700

paasta-tools (1.6.3) xenial; urgency=medium

* 1.6.3 tagged with 'make release'
Commit: Remove vtgate lifecycle hooks (#3966) Remove vtgate
lifecycle hooks as we"re not proceeding with #3959 and would like to
have functional vtgate pods for testing

-- Luis Perez <[email protected]> Mon, 07 Oct 2024 12:24:40 -0700

paasta-tools (1.6.2) xenial; urgency=medium

* 1.6.2 tagged with 'make release'
Commit: Sort dropped caps when caps are added (#3973) As sets are
unordered, not sorting the resulting list built from set operations
means that we"re constantly changing the order of metadata in the
final podspec, leading to bounces almost every time the s_k_j runs

-- Luis Perez <[email protected]> Wed, 25 Sep 2024 15:07:47 -0700

paasta-tools (1.6.1) xenial; urgency=medium

* 1.6.1 tagged with 'make release'
Commit: Only drop capabilities that are not added (#3972) It
appears that containerd has changed the behavior around
adding/dropping linux capabilities and added caps no longer take
precedence over dropped ones

-- Luis Perez <[email protected]> Wed, 25 Sep 2024 09:39:23 -0700

paasta-tools (1.6.0) xenial; urgency=medium

* 1.6.0 tagged with 'make release'
Commit: Add service and instance annotations to tron pods (#3967)
We can currently only figure out what service/instance a log belongs
to by looking at the k8s labels for the emitting pod, but label
values are quite limited in length and we"ve got some pretty large
job and/or action names, which means that the instance label for a
large chunk of tronjobs ends up getting truncated. Solution:
annotations! these have a significantly higher limit (256kb) and
they can still be read by our otel collector - the only downside is
that annotations cannot be used for filtering, but that"s fine :)

-- Luis Perez <[email protected]> Wed, 25 Sep 2024 07:59:52 -0700

paasta-tools (1.5.5) xenial; urgency=medium

* 1.5.5 tagged with 'make release'
Commit: Fixing paasta clean up stale nodes (#3971) * Fixing paasta
clean up stale nodes * Fixing paasta clean up stale nodes

-- Wilmer Bandres <[email protected]> Wed, 25 Sep 2024 04:45:30 -0700

paasta-tools (1.5.4) xenial; urgency=medium

* 1.5.4 tagged with 'make release'
Commit: Document minimum_error_rps in smartstack.yaml (#3970)
Update the paasta docs with minimum_error_rps for default error
alerting endpoints

-- Luis Perez <[email protected]> Tue, 24 Sep 2024 12:09:45 -0700

paasta-tools (1.5.3) xenial; urgency=medium

* 1.5.3 tagged with 'make release'
Commit: Merge pull request #3968 from Yelp/u/mpiano/SEC-19486 local-
run: add option to authenticate as the calling user

-- Eman Elsabban <[email protected]> Mon, 23 Sep 2024 12:41:04 -0700

paasta-tools (1.5.2) xenial; urgency=medium

* 1.5.2 tagged with 'make release'
Commit: Upgrading k8s lib (#3965) * Upgrading k8s lib * Fixing
tests * fixing setup crd * fixing setup CR * Removing dead code

-- Wilmer Bandres <[email protected]> Fri, 20 Sep 2024 08:57:21 -0700

paasta-tools (1.5.1) xenial; urgency=medium

* 1.5.1 tagged with 'make release'
Commit: Merge pull request #3960 from Yelp/yaro/update_vector-
logs_filer Update vector-logs message filter FN usage

-- Yaroslav Liakhovskyi <[email protected]> Wed, 18 Sep 2024 00:46:54 -0700

paasta-tools (1.5.0) xenial; urgency=medium

* 1.5.0 tagged with 'make release'
Commit: Attempt to batch config loading for tron deployments (#3956)
Right now we make at most 2N calls to the Tron API during config
deployments: N to get the current configs and at most N if all
services have changes. To start, I"d like to reduce this to N by
allowing GET /api/config to return all the configs so that the only
requests needed are POSTs for changed configs. Since I"m a little
worried about how tron-pnw-prod will fare with returning all the
configs in a single request, using this new endpoint is behind a
feature toggle. Hopefully this works out and we don"t need to add
pagination to the endpoint :) Depending on how this goes, we can
look into batching up the POSTs so that we can also do that in a
single request (or at least <N requests if we want to chunk things a
bit more).

-- Luis Perez <[email protected]> Tue, 17 Sep 2024 12:46:59 -0700

paasta-tools (1.4.34) xenial; urgency=medium

* 1.4.34 tagged with 'make release'
Commit: Merge pull request #3958 from Yelp/jfong/PAASTA-18194-all-
namespaces-option PAASTA-18194: Stop querying all managed
namespaces by default for paasta status

-- Jen Patague <[email protected]> Mon, 16 Sep 2024 12:11:03 -0700

paasta-tools (1.4.33) xenial; urgency=medium

* 1.4.33 tagged with 'make release'
Expand Down
Loading

0 comments on commit 998333a

Please sign in to comment.