* Avoid database query by passing event directly to job-queue
* Avoid block progress query by returning from update query
* Process batch of events for a block in a single job
* Fix smoke test for subscribed events and use teamSize for job queue
* Perf improvement TODOs
* Move watch contract to util package
* Get block progress from event instead of querying
* Use index field in event to check order of processing
* Use watched contract map to avoid querying database
* Use update query for blockProgress entity
Co-authored-by: Ashwin Phatak <ashwinpphatak@gmail.com>
* Implement clean jobs CLI in watchers
* Pull common code cleanJobs to util
* Implement commands for reset job-queue and state
* Reset command for job-queues and watcher state
* Reset sync status on reset state
* Fix using cli from built js
* Implement skipCheck to reset syncStatus table and skip complete handler on fail
* Check for block isComplete on reset state
* Set default value for force param to update syncStatus to false
* Reset tables in erc20 watcher
* Push job in fill after previos block is complete
* Implement last processed event in uni-info-watcher.
* Check if block or event is already processed.
Co-authored-by: nabarun <nabarun@deepstacksoft.com>
* Handle reorgs upto known canonical block hash.
* Process block events strictly in order.
* Support filling old blocks, process in-order.
* Always publish block progress.
