feat: Serve 10% of tag-page-rendering PROD traffic with ECS

[akash1810]

What does this change?

This change updates the traffic weighting configuration of tag-page-rendering:

• In CODE serve 100% of traffic with ECS
• In PROD serve 10% of traffic with ECS. This is roughly 70-100 requests / minute.

Why?

Though we have load tested (see below), real production traffic is the best way to understand how the ECS infrastructure will perform.

Also, this is a step closer to removing EC2 infrastructure!

How has this change been tested?

In #16693 PROD was configured to run 9 tasks and scale at 20% CPU (up to 90 tasks). With this configuration, we've performed a load test (see #16452) with the following results:

Note

We think PROD ECS can serve 100% of traffic, however to be conservative, we'll start with 10%.

---

Co-authored-by: @connoromalleyatwork

[connoromalleyatwork]

Looks good!!

[github-actions]

Deploy build 3242 of dotcom:ab-testing to CODE

All deployment options

• Deploy build 3242 of dotcom:ab-testing to CODE
• Deploy parts of build 3242 to CODE by previewing it first
• What's on CODE right now?

---

From guardian/actions-riff-raff.

[github-actions]

🚀 Image pushed to AWS ECR

Image digest: sha256:a560de7b6a831fc522dab401b18cc7431ef55c4e0ce3ae4bebfc66ca522c2cfd

🐛 Run the image locally

The following can be used to run the image locally:

# Refer to image using the immutable digest. Find alternatives below.
IMAGE_IDENTIFIER="@sha256:a560de7b6a831fc522dab401b18cc7431ef55c4e0ce3ae4bebfc66ca522c2cfd"

# Refer to image using branch tag
# IMAGE_IDENTIFIER=":branch-ecs-weighting"

# Refer to image using build tag
# IMAGE_IDENTIFIER=":build-31292"

# Set environment variables for the AWS CLI
AWS_PROFILE="<A_PROFILE_FROM_JANUS>"
AWS_DEFAULT_REGION="eu-west-1"

IMAGE_ACCOUNT_ID=$(aws ssm get-parameter --name /organisation/accounts/artifacts --query "Parameter.Value" --output text)
REGISTRY="${IMAGE_ACCOUNT_ID}.dkr.ecr.${AWS_DEFAULT_REGION}.amazonaws.com"
IMAGE="${REGISTRY}/guardian/dotcom-rendering${IMAGE_IDENTIFIER}"

# Login to AWS ECR https://docs.aws.amazon.com/AmazonECR/latest/userguide/registry_auth.html
aws ecr get-login-password | docker login --username AWS --password-stdin $REGISTRY

# Pull the image
docker pull $IMAGE

# Run the image. You'll likely need to set additional flags. See https://docs.docker.com/reference/cli/docker/container/run.
docker run $IMAGE

---

From guardian/actions-publish-image.

[github-actions]

Deploy build 31292 of dotcom:rendering-all to CODE

All deployment options

• Deploy build 31292 of dotcom:rendering-all to CODE
• Deploy parts of build 31292 to CODE by previewing it first
• What's on CODE right now?

---

From guardian/actions-riff-raff.

[gu-prout]

Seen on PROD (merged by @akash1810 9 minutes and 9 seconds ago) Please check your changes!

[akash1810]

This is live as of https://riffraff.gutools.co.uk/deployment/view/7824d0b4-1a12-404f-9fb7-5287e6c72194.

Average latency is up very slightly, the availability SLI remains at 100% though:

[akash1810]

Metrics from ECS for the last 12 hours:

Looking pretty good! As a reminder, the EC2 infrastructure scales out at 400ms latency.