As part of elastic/uptime#432 we'll need to support sending zips to GCP cloud storage after the discovery phase, and the retrieval of that same asset in the execution phase. This lets the discovery phase happen on one node, and many parallel invocations of execution happen on other nodes without re-downloading the asset from the original source. Additionally, we can upload a zip with node_modules pre-installed speeding up execution.
This would work hand-in-hand with #29917 and is blocked by it.
Sample Discovery Config
# Discovery config
heartbeat.monitors:
- id: elastic-monitor
name: Elastic Monitor
discovery:
enabled: true
store:
gcp:
bucket: "my-bucket-name"
base_path: "/path/to/my/zips"
auth: # GCP auth settings if not using instance creds
schedule: '@every 1m'
source:
zip_url:
url: "https://github.com/elastic/synthetics-demo/archive/refs/heads/main.zip"Additional fields in discovery doc
The bucket and full path would be added to the discovery doc as:
{
"suite": { }, // omitting for brevity
"cached_source": {
"gcp_url": { // note the structure here is identical to the heartbeat yaml in the execution config under `source`
"bucket": "my-bucket-name",
"path": "/path/to/my/zips/oeuh89g89h9823.zip"
}
}
}Sample Execution Config
# Execution config
heartbeat.monitors:
- id: elastic-monitor
name: Elastic Monitor
schedule: '@every 1m'
source:
gcp_url:
bucket: "my-bucket-name"
path: "/path/to/my/zips/oeuh89g89h9823.zip"
filter_journeys:
match: my-exact-journey-name
As part of elastic/uptime#432 we'll need to support sending zips to GCP cloud storage after the discovery phase, and the retrieval of that same asset in the execution phase. This lets the discovery phase happen on one node, and many parallel invocations of execution happen on other nodes without re-downloading the asset from the original source. Additionally, we can upload a zip with
node_modulespre-installed speeding up execution.This would work hand-in-hand with #29917 and is blocked by it.
Sample Discovery Config
Additional fields in discovery doc
The bucket and full path would be added to the discovery doc as:
Sample Execution Config