Commit graph

32 commits

Author SHA1 Message Date
Simon Mo
8beb887bbe
[Serve] Remove release tests for checkpoint_path (#27194) 2022-07-28 12:30:30 -07:00
Simon Mo
e5a8b1dd55
[Serve] Add API Annotations And Move to _private (#27058) 2022-07-27 09:08:26 -07:00
Sihan Wang
8ecd928c34
[Serve] Make the checkpoint and recover only from GCS (#26753) 2022-07-25 14:24:53 -07:00
Archit Kulkarni
e043f49957
[Serve] [CI] Increase instance size and add debug log for autoscaling_multi_deployment release test (#26732) 2022-07-20 16:13:36 -07:00
Sihan Wang
b606169cb5
[Serve] Promote autoscaling feature (#26393)
1. get rid of the private attribute
2. fix unit test
3. docs and workflows
2022-07-13 14:38:38 -05:00
Simon Mo
438b6c78c8
[Release Tests] Add memory monitoring for Serve release test (#25868) 2022-06-17 11:11:56 -07:00
Jiao
f6735f90c7
[Ray DAG] Move dag project folder out of experimental (#25532) 2022-06-16 19:15:39 -07:00
Sihan Wang
b024a9543e
[Serve] Support scale replica down to 0 (#24892) 2022-06-02 08:06:46 -07:00
Sihan Wang
3f5da8af7a
[Serve] Add serve handle graph workload nightly tests (#24435) 2022-05-04 09:07:50 -07:00
Jiao
9d31f5f7b2
[Serve] Change deployment graph long chain test (#24418) 2022-05-03 10:38:47 -07:00
Jiao
ba7cc1803a
[Deployment Graph] Add release test for long chain & wide fanout pattern (#24246) 2022-04-29 17:03:33 -07:00
shrekris-anyscale
b51d0aa8b1
[serve] Introduce context.py and client.py (#24067)
Serve stores context state, including the `_INTERNAL_REPLICA_CONTEXT` and the `_global_client` in `api.py`. However, these data structures are referenced throughout the codebase, causing circular dependencies. This change introduces two new files:

* `context.py`
    * Intended to expose process-wide state to internal Serve code as well as `api.py`
    * Stores the `_INTERNAL_REPLICA_CONTEXT` and the `_global_client` global variables
* `client.py`
    * Stores the definition for the Serve `Client` object, now called the `ServeControllerClient`
2022-04-21 18:35:09 -05:00
Edward Oakes
de227ac407
[serve] Add component logger + basic access logging (#23558)
Adds a "component logger" to standardize logging across the HTTP proxy, controller, and deployment replicas.
2022-04-12 18:16:58 -05:00
Archit Kulkarni
db2c37c760
[serve] [release] Disable smoke test by default (#23334) 2022-03-18 18:40:48 -05:00
Edward Oakes
135cd121b9
[release tests] Fix minor bug in multi-deployment serve test (#22961) 2022-03-09 14:37:27 -06:00
Edward Oakes
aa907987bf
[serve][release tests] Use m5.8xlarge instance types for 1k replica tests (#22918) 2022-03-08 21:34:01 -06:00
Archit Kulkarni
31332f8930
[serve] [release tests] Add health check grace period for 1k deployment (#22651) 2022-02-25 12:13:44 -06:00
Balaji Veeramani
7f1bacc7dc
[CI] Format Python code with Black (#21975)
See #21316 and #21311 for the motivation behind these changes.
2022-01-29 18:41:57 -08:00
shrekris-anyscale
75b3080834
[Serve] Serve Autoscaling Release tests (#21208) 2022-01-21 12:08:25 -08:00
Simon Mo
c85e9e69b3
[Serve] Change multi_deployment_1k_noop_replica threshold (#20514) 2021-11-17 17:25:54 -08:00
Simon Mo
b6bd4fd5f3
[Serve] Don't recover from current state checkpoint (#19998) 2021-11-12 09:02:27 -08:00
Tobias Kaymak
893f57591d
[serve] Add Google Cloud Storage as a backend (#20104) 2021-11-10 19:45:19 -08:00
Simon Mo
4d583da7d5
[Serve] Add verbose log for nightly test only (#20088) 2021-11-04 16:15:22 -07:00
Jiao
3f628d4f6b
increase long poll timeout and wrk trial cpu resource (#19768) 2021-10-26 21:31:39 -07:00
Jiao
85b8a6de5f
[Serve] Add nightly test for Serve failure recovery (#19125) 2021-10-11 18:33:20 -07:00
Jiao
b52c873027
[serve] Use list_deployments in benchmark (#18050) 2021-08-25 12:26:46 -05:00
Jiao
3c64a1a3c1
Add micro benchmark to releaser repo (#17727) 2021-08-11 15:15:33 -07:00
Jiao
2618236167
[serve] Fix single deployment nightly test (#17368) 2021-07-28 11:38:06 -05:00
Jiao
9eb1bcd061
[serve] Multi & single deployment large scale test (#17310) 2021-07-27 10:46:45 -05:00
Edward Oakes
58423e6018
[serve] Improve nightly release test (#17277) 2021-07-26 11:15:46 -05:00
Jiao
7473f663ef
[Release] change replica to 100 to collect signals now (#17214)
Co-authored-by: Jiao Dong <jiaodong@anyscale.com>
2021-07-20 12:27:56 -07:00
Jiao
994ff3ce21
[Serve] Add initial large scale tests (#17026) 2021-07-20 08:56:29 -07:00