Collect telemetry data

Execute up to 1000 telemetry queries. A leaky bucket rate limiting mechanism is applied based on the calculated cost of each request. The cost is a linear combination of the number of requests, the number of queries per request, the estimated number of returned datapoints and the number of scanned shards.

To minimize cost:

• Requests: Send multiple queries in a single request (up to 1000). Make sure the total cost does not exceed your tier's bucket capacity.
• Queries per request: Use a single query to select multiple fields instead of one query per field.
• Returned datapoints: The estimate is based on the selected time range, the expected datapoint frequency and the provided limit. When no range is provided, a worst-case range is assumed. Always specify a time range and set limit as low as possible.
• Scanned shards: Depends on the time range and expected datapoint frequency, but is unaffected by limit. A tight time range minimizes this component as well.

If a request's cost would overflow the bucket, a 429 response is returned with a Retry-After header indicating how long to wait. If the cost exceeds the bucket's total capacity entirely, a 422 is returned. Ignoring Retry-After headers will result in a temporary 5-minute IP ban.

Helper scripts for estimating request cost are available at /assets/estimateCost.ts (TypeScript) and /assets/estimateCost.py (Python).