Rate limits should be definable by extraction-worker pool

[TimDaub]

• many cloud providers have exact rate limits usually denoted in calls/seconds, e.g. https://docs.infura.io/infura/networks/ipfs/how-to/request-rate-limits "You can retrieve IPFS data using the API with a limit of 100 requests per second for the remaining read-only methods such as:"
• But don't confuse this with extraction-worker's concurrency which just says: independent of time, there should always been 100 connections open (we call this concurrency).
• While concurrency is important and we should keep it, another useful option would be to allow users to define rate limit quotas, e.g. "1/s" (1 call per second) or "10/ms" (10 calls per millisecond).
• It'd be ideal if a strategy implementer would add those quotas as property to a worker message: https://github.com/neume-network/message-schema/blob/a66dbf12a37709177aa238b2c8e08ec105a90097/src/index.mjs#L140
• and then extraction-worker would queue those calls respectively such that they never go over the defined quota, e.g. extraction-worker would guarantee that many messages with a quota: "1/s" would never exceed 1 call per second etc.

[TimDaub]

closed in #38