Monitoring Cloudflare Worker Performance with Datadog

Question

Pulumi · Accepted Answer

To set up monitoring for Cloudflare Worker performance with Datadog, you will need to create a monitor in Datadog that tracks specific metrics relevant to Cloudflare Workers. Typically, you might track metrics like request count, error rate, or latency. These metrics can either be pushed to Datadog using the DogStatsD agent or fetched by Datadog via integrations or the API, depending on the capabilities of Cloudflare's monitoring tools and how they can interoperate with Datadog.

In Pulumi, you do this by creating a `datadog.Monitor` resource. This resource will define the type of monitor (e.g., metric alert, anomaly detection), the query for the metric you want to track, a message detailing what the monitor is for and steps to take in case of an alert, and other features like tags for organizational purposes.

Please note that while Pulumi can help you create the infrastructure to monitor your Cloudflare Workers from Datadog's side, you will also need to make sure that the relevant metrics are being sent to Datadog from Cloudflare Workers. This might involve configuring additional Cloudflare or Datadog settings outside of Pulumi.

Here is a Python program to create a Datadog monitor for tracking the average latency of a Cloudflare Worker. The program assumes that you have already configured Pulumi for use with the Datadog provider:

```python
import pulumi
import pulumi_datadog as datadog

# Example: Creating a Datadog monitor to alert on high latency in a Cloudflare Worker.
# In this case, "cloudflare.worker.latency" is the metric to be monitored.
# Replace it with the actual metric you want to monitor from Cloudflare Workers.
cloudflare_latency_monitor = datadog.Monitor("cloudflareLatencyMonitor",
    type="query alert",  # This is a query alert type, which will trigger based on the result of the query
    query="avg(last_5m):avg:cloudflare.worker.latency{environment:production} > 100",  # Modify accordingly
    message="""\
        Team, the Cloudflare Worker has a high latency issue!
        @pagerduty - Take immediate action.
        @webhook - Alerting our systems.
    """,  # This message will be sent when the alert triggers
    tags=["cloudflare", "performance", "latency"],  # Useful for organizing and filtering monitors
    priority=3,  # An integer from 1 (high) to 5 (low) indicating alert severity
)

pulumi.export("cloudflare_latency_monitor_id", cloudflare_latency_monitor.id)  # Exports the monitor ID
```

In the program, we first import the required Pulumi and Datadog packages. Next, we create a `datadog.Monitor` object with a query that alerts if the average latency of a Cloudflare Worker exceeds 100 milliseconds in the last five minutes. The query will need to be adjusted according to your actual Datadog metric name and conditions.

The `message` attribute is where you would specify the text of the alert, including any directives to alerting channels like PagerDuty or webhooks.

Lastly, we export the monitor ID so that you can easily reference it outside the program, perhaps in other Pulumi stacks or in your CI/CD pipeline.

Make sure that metrics are appropriately tagged in Datadog and that your query filters match those tags to monitor the right data.

To be able to run this code you need to install the `pulumi_datadog` package, set up your Pulumi stack, and [configure your Datadog provider credentials](https://www.pulumi.com/registry/packages/datadog/installation-configuration/). The necessary configurations are Datadog API key and application key which will be used to authenticate API requests.