KNOX-2746 - Add presto/presto ui support in service definition

[Bhargavi-Sagi]

What changes were proposed in this pull request?

Presto is a distributed SQL query engine designed to query large data sets distributed over one or more heterogeneous data sources. Add support for in knox service definition, so that knox can provide proxying service for presto.

How was this patch tested?

Deployed knox in a cluster with presto, and checked access presto ui with knox server port 8442. Also checked presto apis with knox server as proxy

API testing:

curl --insecure -X POST -d 'show catalogs' -H "X-Presto-User: user" -H "Content-Type: application/json" https://localhost:8442/gateway/default/presto/v1/statement

{"id":"20220517_112614_00000_muw7k","infoUri":"https://localhost:8442/gateway/default/presto/ui/query.html?20220517_112614_00000_muw7k","nextUri":"https://localhost:8442/gateway/default/presto/v1/statement/queued/20220517_112614_00000_muw7k/1?slug=x31aa07b4664b44de90fe15c2e2a758cc","stats":{"state":"WAITING_FOR_PREREQUISITES","waitingForPrerequisites":true,"queued":false,"scheduled":false,"nodes":0,"totalSplits":0,"queuedSplits":0,"runningSplits":0,"completedSplits":0,"cpuTimeMillis":0,"wallTimeMillis":0,"waitingForPrerequisitesTimeMillis":0,"queuedTimeMillis":0,"elapsedTimeMillis":0,"processedRows":0,"processedBytes":0,"peakMemoryBytes":0,"peakTotalMemoryBytes":0,"peakTaskTotalMemoryBytes":0,"spilledBytes":0},"warnings":[]}

curl --insecure https://localhost:8442/gateway/default/presto/v1/statement/queued/20220517_112614_00000_muw7k/1?slug=x31aa07b4664b44de90fe15c2e2a758cc

{"id":"20220517_112614_00000_muw7k","infoUri":"https://localhost:8442/gateway/default/presto/ui/query.html?20220517_112614_00000_muw7k","nextUri":"https://localhost:8442/gateway/default/presto/v1/statement/queued/20220517_112614_00000_muw7k/2?slug=x31aa07b4664b44de90fe15c2e2a758cc","stats":{"state":"WAITING_FOR_PREREQUISITES","waitingForPrerequisites":true,"queued":false,"scheduled":false,"nodes":0,"totalSplits":0,"queuedSplits":0,"runningSplits":0,"completedSplits":0,"cpuTimeMillis":0,"wallTimeMillis":0,"waitingForPrerequisitesTimeMillis":6,"queuedTimeMillis":0,"elapsedTimeMillis":6,"processedRows":0,"processedBytes":0,"peakMemoryBytes":0,"peakTotalMemoryBytes":0,"peakTaskTotalMemoryBytes":0,"spilledBytes":0},"warnings":[]}

curl --insecure https://localhost:8442/gateway/default/presto/v1/statement/queued/20220517_112614_00000_muw7k/2?slug=x31aa07b4664b44de90fe15c2e2a758cc

{"id":"20220517_112614_00000_muw7k","infoUri":"https://localhost:8442/gateway/default/presto/ui/query.html?20220517_112614_00000_muw7k","partialCancelUri":"https://localhost:8442/gateway/default/presto/v1/stage/20220517_112614_00000_muw7k.0","nextUri":"https://localhost:8442/gateway/default/presto/v1/statement/executing/20220517_112614_00000_muw7k/1?slug=x31aa07b4664b44de90fe15c2e2a758cc","columns":[{"name":"Catalog","type":"varchar(14)","typeSignature":{"rawType":"varchar","typeArguments":[],"literalArguments":[],"arguments":[{"kind":"LONG_LITERAL","value":14}]}}],"data":[["awsdatacatalog"],["hive"],["system"]],"stats":{"state":"RUNNING","waitingForPrerequisites":false,"queued":false,"scheduled":true,"nodes":1,"totalSplits":19,"queuedSplits":0,"runningSplits":0,"completedSplits":19,"cpuTimeMillis":14,"wallTimeMillis":141,"waitingForPrerequisitesTimeMillis":15,"queuedTimeMillis":0,"elapsedTimeMillis":12522,"processedRows":0,"processedBytes":0,"peakMemoryBytes":0,"peakTotalMemoryBytes":171,"peakTaskTotalMemoryBytes":171,"spilledBytes":0,"rootStage":{"stageId":"0","state":"RUNNING","done":false,"nodes":1,"totalSplits":1,"queuedSplits":0,"runningSplits":0,"completedSplits":1,"cpuTimeMillis":1,"wallTimeMillis":3,"processedRows":3,"processedBytes":67,"subStages":[{"stageId":"1","state":"FINISHED","done":true,"nodes":1,"totalSplits":17,"queuedSplits":0,"runningSplits":0,"completedSplits":17,"cpuTimeMillis":13,"wallTimeMillis":138,"processedRows":3,"processedBytes":67,"subStages":[{"stageId":"2","state":"FINISHED","done":true,"nodes":1,"totalSplits":1,"queuedSplits":0,"runningSplits":0,"completedSplits":1,"cpuTimeMillis":0,"wallTimeMillis":0,"processedRows":3,"processedBytes":0,"subStages":[]}]}]},"runtimeStats":{"S1-driverCountPerTask":{"name":"S1-driverCountPerTask","sum":17,"count":1,"max":17,"min":17},"S2-driverCountPerTask":{"name":"S2-driverCountPerTask","sum":1,"count":1,"max":1,"min":1},"S1-taskElapsedTimeNanos":{"name":"S1-taskElapsedTimeNanos","sum":142799852,"count":1,"max":142799852,"min":142799852},"S2-taskElapsedTimeNanos":{"name":"S2-taskElapsedTimeNanos","sum":106038415,"count":1,"max":106038415,"min":106038415},"S0-driverCountPerTask":{"name":"S0-driverCountPerTask","sum":1,"count":1,"max":1,"min":1},"S0-taskElapsedTimeNanos":{"name":"S0-taskElapsedTimeNanos","sum":0,"count":1,"max":0,"min":0}},"progressPercentage":100.0},"warnings":[]}

[smolnar82]

@Bhargavi-Sagi - this looks great!
There was another PR recently that added Trino as a new service to Knox. Checking Trino's and Presto's pages I found exactly the same description of the 2 projects:


On Trino's UI there's even a link suggesting we are talking about the same (similar) services. Hence the question: do we need two separate service definitions for the same functionality?

[Bhargavi-Sagi]

Presto and Trino are different projects. Trino was initially forked from presto but then they have diverged quite a lot over time. There are apis in Trino which are not present in presto. Even though UI page looks same not all the apis are same(for example login/logout are not present in presto) and so we will need separate service definitions for Presto and Trino.

[smolnar82]

LGTM