issue_comments

thanks @simonw!

Tested that with:

datasette publish cloudrun fixtures.db --service issue-1779 --min-instances 2 --max-instances 4

Here's the start of the man page for gcloud run deploy: ``` NAME gcloud run deploy - deploy a container to Cloud Run

SYNOPSIS gcloud run deploy [[SERVICE] --namespace=NAMESPACE] [--args=[ARG,...]] [--async] [--command=[COMMAND,...]] [--concurrency=CONCURRENCY] [--cpu=CPU] [--ingress=INGRESS; default="all"] [--max-instances=MAX_INSTANCES] [--memory=MEMORY] [--min-instances=MIN_INSTANCES] `` I'm going to expose--max-instancesand--min-instancesas extra option todatasette publish cloudrun`.

(I deleted my issue-1779 project using the UI at https://console.cloud.google.com/run?project=datasette-222320)

Just tried this instead, and it still worked and deployed OK:

datasette publish cloudrun fixtures.db --memory 16Gi --cpu 4 --service issue-1779

@fgregg I'm not able to replicate your deployment failure I'm afraid.

Just spotted this in the failing Actions workflow:

gcloud config set run/region us-central1

I tried that locally too but the deploy still succeeds.

datasette publish cloudrun fixtures.db --memory 8Gi --cpu 2 --service issue-1779 still works.

Maybe I need to upgrade: % gcloud --version Google Cloud SDK 378.0.0 alpha 2022.03.18 bq 2.0.74 core 2022.03.18 gsutil 5.8 Updates are available for some Google Cloud CLI components. To install them, please run: $ gcloud components update ``` % gcloud components update Beginning update. This process may take several minutes.

Your current Google Cloud CLI version is: 378.0.0 You will be upgraded to version: 397.0.0

┌─────────────────────────────────────────────────────────────────────────────┐ │ These components will be updated. │ ├─────────────────────────────────────────────────────┬────────────┬──────────┤ │ Name │ Version │ Size │ ├─────────────────────────────────────────────────────┼────────────┼──────────┤ │ BigQuery Command Line Tool │ 2.0.75 │ 1.6 MiB │ │ BigQuery Command Line Tool (Platform Specific) │ 2.0.75 │ < 1 MiB │ │ Cloud Storage Command Line Tool │ 5.11 │ 15.5 MiB │ │ Cloud Storage Command Line Tool (Platform Specific) │ 5.11 │ < 1 MiB │ │ Google Cloud CLI Core Libraries │ 2022.08.05 │ 24.3 MiB │ │ Google Cloud CLI Core Libraries (Platform Specific) │ 2022.08.05 │ < 1 MiB │ │ anthoscli │ 0.2.28 │ 48.0 MiB │ │ gcloud Alpha Commands │ 2022.08.05 │ < 1 MiB │ │ gcloud cli dependencies │ 2022.07.29 │ 11.2 MiB │ └─────────────────────────────────────────────────────┴────────────┴──────────┘

A lot has changed since your last upgrade. For the latest full release notes, please visit: https://cloud.google.com/sdk/release_notes % gcloud --version
Google Cloud SDK 397.0.0 alpha 2022.08.05 bq 2.0.75 core 2022.08.05 gsutil 5.11 ```

Tried duplicating this error locally but the following command succeeded when I expected it to fail:

datasette publish cloudrun fixtures.db --memory 8Gi --cpu 2 --service issue-1779

Here's the relevant part of the datasette publish command from that failed Actions workflow: datasette publish cloudrun f7.db nlrb.db opdr.db old_nlrb.db voluntary_recognitions.db work_stoppages.db lm20.db chips.db \ --memory 8Gi \ --cpu 2

Actually I disagree that Datasette isn't setup to make use of container scaling: part of the idea behind the Baked Data pattern is that you can scale to handle effectively unlimited traffic by running multiple copies of your application, each with their own duplicate copy of the database.

So I'm going to default maxScale to 10 and still let people customize it.

I'm going to default maxscale to 2 but provide an extra command line option for datasette publish cloudrun which lets people set it to something higher if they want to.

maybe a simpler solution is to set the maxscale to like 2? since datasette is not set up to make use of container scaling anyway?