Instana

- Implement batching for very short (< 10 ms), high-frequency database spans (opt-in in this release, will be switched to default behaviour in one of the next releases). To enabled it right now, any of the three following methods can be used:
- Set the environment variable `INSTANA_SPANBATCHING_ENABLED=true`.
- Use in-code configuration: `config.tracing.spanBatchingEnabled: true`.
- Add this to the agent's `configuration.yaml`:


com.instana.plugin.nodejs:
span-batching-enabled: true

- [AWS Lambda] Check type of Lambda error messages and stringify if necessary.

- Add instrumentation for [`google-cloud/pubsub`](https://www.npmjs.com/package/google-cloud/pubsub).

- Fix (Elasticsearch): Adapt instrumentation for `elastic/elasticsearch` for refactoring of `api` submodule introduced in `elastic/elasticsearch7.9.1`. (Fixes `TypeError: Cannot set property 'Symbol(configuration error)' of undefined`.) Also, remove obsolete `stats` annotation for `elasticsearch` spans.
- Remove longer span reporting interval for serverless tracers, use standard interval instead.

- [Google Cloud Run]: Minor changes for the upcoming Google Cloud Run support. Requires at least Instana back end version 189.

- Remove direct `node-fetch` dependency from `instana/aws-fargate` as well as `instana/google-cloud-run` and move it to `instana/metrics-util` instead.
- Optimization: Only set `span.data.service` on first span.
- Send `span.crid`/`span.crtp` instead of `span.data.correlationId`/`span.data.correlationType`.

- Fix: Move `event-loop-lag` from optionalDependencies to dependencies, so installations with `--ignore-optional` will work correctly. ([258](https://github.com/instana/nodejs/pull/258), thanks to ThisIsMissEm).
- [AWS Fargate]: Update dependency `node-fetch` version 2.6.1 (fixes https://www.npmjs.com/advisories/1556).