Add TPU Capacity Validation DAG - #191
Open
chengpinglin wants to merge 7 commits into
Open
Conversation
chengpinglin
requested review from
alfredyu-cienet,
andrewyct,
severus-ho and
xibinliu
as code owners
January 27, 2026 03:16
chengpinglin
changed the title
| "gs://ml-auto-solutions-dag-configs/tpu_observability/dag_config.yaml" | ||
| ) | ||
|
|
||
| VM_GCS_CONFIG_PATH = "gs://ml-auto-solutions-dag-configs/tpu_observability/dag_config_tony_vm.yaml" |
Collaborator
There was a problem hiding this comment.
is this your personal config?
alfredyu-cienet
force-pushed
the
dev
branch
2 times, most recently
from
January 28, 2026 06:23
0f0e4e2 to
31723ef
Compare
chengpinglin
force-pushed
the
tpu-obs/user/tony/capacity_01_dipstick
branch
from
January 28, 2026 07:02
206efaa to
91e3a7d
Compare
alfredyu-cienet
force-pushed
the
dev
branch
2 times, most recently
from
February 2, 2026 08:04
c846590 to
395bf7c
Compare
…tation and description
… TPU chips validation DAG
…d latency calculation
chengpinglin
force-pushed
the
tpu-obs/user/tony/capacity_01_dipstick
branch
from
February 3, 2026 05:45
91e3a7d to
5528e72
Compare
alfredyu-cienet
force-pushed
the
dev
branch
5 times, most recently
from
February 13, 2026 08:57
ac9e9e7 to
0cf6e23
Compare
alfredyu-cienet
force-pushed
the
dev
branch
2 times, most recently
from
March 6, 2026 08:10
e1c48aa to
821fb36
Compare
alfredyu-cienet
force-pushed
the
dev
branch
3 times, most recently
from
May 19, 2026 03:54
8eb5556 to
acf36a1
Compare
alfredyu-cienet
force-pushed
the
dev
branch
3 times, most recently
from
May 28, 2026 07:44
83f8b37 to
ea29ef3
Compare
alfredyu-cienet
force-pushed
the
dev
branch
4 times, most recently
from
June 8, 2026 08:58
4af284a to
a9fa8fb
Compare
alfredyu-cienet
force-pushed
the
dev
branch
4 times, most recently
from
June 17, 2026 08:08
f8824db to
09c9e42
Compare
alfredyu-cienet
force-pushed
the
dev
branch
5 times, most recently
from
July 3, 2026 08:25
d242c8d to
f76e189
Compare
alfredyu-cienet
force-pushed
the
dev
branch
4 times, most recently
from
July 14, 2026 03:12
a577785 to
71b8d8a
Compare
alfredyu-cienet
force-pushed
the
dev
branch
4 times, most recently
from
July 23, 2026 01:43
b588728 to
0e938b1
Compare
alfredyu-cienet
force-pushed
the
dev
branch
2 times, most recently
from
July 27, 2026 07:46
54f9a78 to
8e4ea82
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
This PR introduces the chips_capacity_validation_dag, an automated workflow designed to validate Cloud TPU capacity and health. The DAG automates the lifecycle of a TPU workload—from infrastructure provisioning to deep metric-based health checks—ensuring that TPU chips are not just allocated, but fully functional and utilized.
Introduces four specialized sensors that poll Cloud Monitoring to verify the state of the hardware:
Scheduled Chips: Confirms the system has successfully assigned TPU resources to the VM.
Active Chips: Validates that the TPU driver is active and reporting to the monitoring backend.
Chip State (Health): Specifically checks for the HEALTHY state label to ensure no hardware defects are present.
Utilized Chips: Confirms that the workload (JAX/PyTorch) has successfully initialized the device and is performing computation.
Tests
cloud-ml-auto-solutions)2.13.1Checklist
Before submitting this PR, please make sure (put X in square brackets):