Summary:
we recently changed our image path to be public for release viziers, and private for rc viziers.
that broke the updater job, because cloudconnector had the private path hardcoded. i made another change to use the path from the current cloudconnector image, but that was incorrect because it doesn't allow you to update from an rc -> public release and vice versa.
that was mostly because i didnt want to hardcode the image paths, but it looks like we'll need to. we should decide what the image path is based on whether the image tag has a "-" in it, which indicates that it is an rc

Test Plan: created an rc, tried to update it to another rc and tried to update it to an official release

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5980

GitOrigin-RevId: d80edcf2cb7403da541bb02cb0e9abe6961c245f

Summary:
Reduce the number of SSL handshakes and add keep alives. This should help a lot with auto complete latencies.

    Original:

    ```
    [0] % hey -c 100 -z 5s https://work.staging.withpixie.dev/api/graphql

    Summary:
      Total:	5.3929 secs
      Slowest:	2.7450 secs
      Fastest:	0.0381 secs
      Average:	0.5370 secs
      Requests/sec:	182.8340

      Total data:	80852 bytes
      Size/request:	82 bytes

    Response time histogram:
      0.038 [1]	|
      0.309 [580]	|■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■
      0.579 [82]	|■■■■■■
      0.850 [31]	|■■
      1.121 [72]	|■■■■■
      1.392 [69]	|■■■■■
      1.662 [70]	|■■■■■
      1.933 [22]	|■■
      2.204 [37]	|■■■
      2.474 [15]	|■
      2.745 [7]	|

    Latency distribution:
      10% in 0.0423 secs
      25% in 0.0486 secs
      50% in 0.0832 secs
      75% in 1.0451 secs
      90% in 1.5565 secs
      95% in 2.0342 secs
      99% in 2.3855 secs

    Details (average, fastest, slowest):
      DNS+dialup:	0.0243 secs, 0.0381 secs, 2.7450 secs
      DNS-lookup:	0.0101 secs, 0.0000 secs, 0.1005 secs
      req write:	0.0000 secs, 0.0000 secs, 0.0001 secs
      resp wait:	0.5119 secs, 0.0379 secs, 2.6882 secs
      resp read:	0.0000 secs, 0.0000 secs, 0.0001 secs

    Status code distribution:
      [401]	986 responses

    ```

    New:

    ```

    Summary:
      Total:	5.1055 secs
      Slowest:	0.7370 secs
      Fastest:	0.0252 secs
      Average:	0.0819 secs
      Requests/sec:	1211.4484

      Total data:	507170 bytes
      Size/request:	82 bytes

    Response time histogram:
      0.025 [1]	|
      0.096 [4700]	|■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■
      0.168 [1140]	|■■■■■■■■■■
      0.239 [264]	|■■
      0.310 [27]	|
      0.381 [20]	|
      0.452 [1]	|
      0.523 [1]	|
      0.595 [1]	|
      0.666 [14]	|
      0.737 [16]	|

    Latency distribution:
      10% in 0.0417 secs
      25% in 0.0520 secs
      50% in 0.0659 secs
      75% in 0.0943 secs
      90% in 0.1360 secs
      95% in 0.1750 secs
      99% in 0.2917 secs

    Details (average, fastest, slowest):
      DNS+dialup:	0.0019 secs, 0.0252 secs, 0.7370 secs
      DNS-lookup:	0.0000 secs, 0.0000 secs, 0.0023 secs
      req write:	0.0000 secs, 0.0000 secs, 0.0002 secs
      resp wait:	0.0796 secs, 0.0252 secs, 0.7369 secs
      resp read:	0.0000 secs, 0.0000 secs, 0.0004 secs

    Status code distribution:
      [401]	6185 responses

    ```

Test Plan: Tested on staging

Reviewers: michelle, nserrino, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5982

GitOrigin-RevId: 90f99e177510e4604f210c567fecab4fb6699208

Summary:
So we had an issue where the app kept using the old index.html because of cache policies. The old index loads JS files that no longer exists since they are hashed by the content.
Our nginx config then serves up the index.html b/c of the missing files. This fix causes index.html to never get cached.

Test Plan: Updated staging and verfied that index.html is always fetched.

Reviewers: michelle, nserrino, #engineering

Reviewed By: nserrino, #engineering

JIRA Issues: PC-511

Differential Revision: https://phab.corp.pixielabs.ai/D5971

GitOrigin-RevId: 9880373bb8d4ec28da6ae8b5129f1ccd2a15ae22

Summary:
ASAN test = slow
BPF test = slow
ASAN + BPF test = more slow

Test Plan: None

Reviewers: yzhao, zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5972

GitOrigin-RevId: b5059278518e6c18dc746a1b6627a3c3c30b1486

Summary: Consistency. We agreed to lowercase p.

Test Plan: Jenkins

Reviewers: yzhao, zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5973

GitOrigin-RevId: 000cde9ca23b4bf430ece01731a1cd56681226e4

Summary:
The initial time the update job applies the bootstrap YAML is so that the clusterroles get updated before running the following steps which may depend on the new clusterrole.
however, if something in the bootstrap YAML requires the updated clusterrole (such as podsecuritypolicy), depending on the ordering of how the YAML is applied, things may fail.
updated the apply YAML code so that we can apply only certain resourceTypes within that YAML. this will allow us to only apply the new clusterrole initially.
the rest of the bootstrap YAML will be applied at a later point, except now with the guarantee that the clusterrole is updated.

Test Plan: updated a vizier

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5970

GitOrigin-RevId: 3b3e7a9a15d045ba3b52db8b7798cf4a139c55b4

Summary: tsia.

Test Plan: updated unit

Reviewers: zasgar, michelle, #engineering

Reviewed By: zasgar, michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5967

GitOrigin-RevId: 6c323a340d4c3f3ed1dfff823e2de4080f6f1027

Summary: Tracing function latency without any explicit arguments would fail to create the argstash, resulting in a failure.

Test Plan: Added tests.

Reviewers: michelle, yzhao, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5969

GitOrigin-RevId: 383ded368108e518e4eb93829ee985ed76c32ca7

Summary:
implemented the part of the vizier updater which actually adds viziers on a queue and processes the queue.
this part is still not yet integrated into the vzmgr logic yet, so things will stop operator as normal after this diff lands. the actual changes  will come in the next diff after things are more tested.

Test Plan: unit tests

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5968

GitOrigin-RevId: 0620e070edc92fc60f7d6780760683d49c011b55

Summary:
we are planning to add an auto-updater in vzmgr.
before we can do that, there needs to be some cleanup done. i pulled out the logic for updating viziers into a new struct, which will be responsible for doing the auto-update in the future.

Test Plan: deployed to staging and deployed a new cluster + updated, to make sure it still works.

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5964

GitOrigin-RevId: 921ec19b1043820bfb0759749c9d2536a362dec7

Summary:
we were occasionally running into prototool errors complaining about the walk-timeout:
```
walking the directory structure looking for proto files timed out after 3s and having seen 20428 files, are you sure you are operating in the right context?
 UNIT ERRORS  Unit testing raised errors!
```

We had run into this issue previously, but the walk-timeout flag hadn't been released yet, so we did a temporary fix: https://phab.corp.pixielabs.ai/D4562
However, there was a release with the flag as of May 19, so we are updating to that version

Test Plan: ran docker image, check that the new version of prototool runs properly with --walk-timeout

Reviewers: oazizi, zasgar, #engineering

Reviewed By: oazizi, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5959

GitOrigin-RevId: 3a5362ed50dac35f1f8c846920ba5f32f2a590be

Summary:
In the new end-to-end streaming Vizier, GRPC sinks will stream results directly to the query broker, rather than Kelvin buffering up the final results and sending it in batch to the query broker. That means there are two 'types' of GRPC sinks in that system: internal GRPC sinks which send mid-query, intermediate data to GRPC sources on another Carnot instance, and external GRPC sinks which send complete results to the query broker or another external address. In the internal GRPC Sink case, the node only needs to know the destination ID of the GRPC Source node that it's sending the data to. In the external GRPC Sink case, the node needs to know the name and schema of the output table.

In this diff, the concept of the external GRPC sink is introduced. Changes to rules are made so that things like automatically adding a limit to memory sinks will also apply to these external GRPC sinks. Next diff, the compiler will change so that px.display automatically results in these external GRPC sinks rather than memory sinks.

Test Plan: added / existing

Reviewers: philkuz, zasgar, #engineering

Reviewed By: philkuz, #engineering

JIRA Issues: PP-2117

Differential Revision: https://phab.corp.pixielabs.ai/D5966

GitOrigin-RevId: 472d7f33454b3b1df3ee5183e80cf62966de1d6f

Summary: Target name and output shouldn't have the same name.

Test Plan: Jenkins

Reviewers: yzhao, #engineering, zasgar

Reviewed By: #engineering, zasgar

Differential Revision: https://phab.corp.pixielabs.ai/D5957

GitOrigin-RevId: 24b8f9c2b2652ba11f0e77ec167846b6dc30c226

PP-2117: When Kelvin is executing a streaming plan, don't call the batch API when the query completes.

Summary:
Currently, there are two APIs for Kelvin sending results to the query broker: TransferResultChunk and ReceiveAgentQueryResult. TransferResultChunk is streaming and will subsume ReceiveAgentQueryResult, which is batch. The query broker has changes to receive results from both, but this Kelvin change is needed so that we only call the ReceiveAgentQueryResult API in situations where the Kelvin plan produces a memory sink.

ReceiveAgentQueryResult reads memory sinks, but in the new streaming API, the sinks will send the data directly to the query broker as GRPC sinks. As a result, during the temporary phase where we support both APIs, we differentiate between which API to use to send the results based on whether or not there are memory sinks in the final plan.

Afterwards, ReceiveAgentQueryResult will be fully deprecated and a lot of this code will go away.

Also, deprecated query_str fully from the execute query request, because it is no longer used anywhere.

Test Plan: added unit test

Reviewers: zasgar, michelle, philkuz, #engineering

Reviewed By: zasgar, #engineering

JIRA Issues: PP-2117

Differential Revision: https://phab.corp.pixielabs.ai/D5961

GitOrigin-RevId: ae680d647a895adbc965e8b2b36c693534bb9656

Summary: tsia

Test Plan: unit tests

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5951

GitOrigin-RevId: 3665753e5eec39bc60b6ed6d2d86a36ea77bfa56

Summary:
Made some tweaks to the way that the output schema will be gotten from the query.
TransferResultChunkRequest will no longer send the output schema (this field was never hooked up to anything).
Instead, the GRPC Sinks that are not internal sinks will contain the full information about their output relation,
and those will be walked and collected (alongside memory sinks) and passed to the client using the existing code paths.
Needed to write a utility to convert the plan map to the output schema, so added a unit test for that as well.
A lot of this logic will be removed once ReceiveAgentResult is deprecated, but OutputSchemaFromPlan should continue to
be necessary and stick around.

Test Plan: unit test

Reviewers: zasgar, philkuz, michelle, #engineering

Reviewed By: zasgar, #engineering

JIRA Issues: PP-2117

Differential Revision: https://phab.corp.pixielabs.ai/D5960

GitOrigin-RevId: bee6d44fbe69e6788dec811d21c0b9f18da5c919

Summary: tsia

Test Plan: webpack

Reviewers: zasgar, philkuz, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5958

GitOrigin-RevId: a4c3eec76fb1e32dbc5df71802b08ebfb2bca9fa

Summary:
Use LaunchDarkly to feature flag the new autocomplete.
The LaunchDarkly feature flag is called 'new-auto-complete' in the LD UI.

Test Plan: Ran webpack and verified manually

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5949

GitOrigin-RevId: 6fbfb697189208eb66489af97b65fd4c47c5a469

Summary: we want to use launchdarkly to control feature flags. this diff adds it to our UI

Test Plan: ran webpack + dev cloud to make sure correct Client ID is used and that my user shows up in the LD UI

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5948

GitOrigin-RevId: 2876e77fe9e05cfe093494cb44b423dabd9a78a1

Summary:
This diff creates some temporary glue code to support the new result streamign API, TransferResultChunk.
Until I deprecate its batch predecessor, ReceiveAgentQueryResult, we still need to rely on the query executor to receive and batch
up results from TransferResultChunk, because the query executor currently expects batch results from ReceiveAgentQueryResults.
As a result, this diff will be another step close to full streaming from Kelvin to client, but it continues to batch results in order to
stay compatible with the existing API. Next step will be to remove ReceiveAgentQueryResults from Kelvin plans. At that point, a lot of
this diff's temporary code will be refactored/removed, and the batch-oriented logic in query executor will be subsumed by new stream-oriented
logic in query result forwarded (which remains a dumb-ish placeholder api for now).

Depends on D5944.

Test Plan: added unit test to query executor to ensure results are stitched correctly.

Reviewers: michelle, zasgar, philkuz, #engineering

Reviewed By: zasgar, #engineering

JIRA Issues: PP-2117

Differential Revision: https://phab.corp.pixielabs.ai/D5954

GitOrigin-RevId: 813d9c0fda99059501a7b44e3a227138cdc9c800

Summary:
This is the endpoint that will be responsible for receiving result chunks from the agent.
It will subsume some of the logic in the ExecuteScript endpoint that is responsible for waiting for query results.

I moved mockContext to be a general testing utility (previously used only in one test).

Also, there was a name collision in the protobuf api, there were two objects being produced both called ExecutionAndTimingInfo (one with an underscore at the end).
Since this field isn't currently used anywhere, I renamed the message type to QueryExecutionAndTimingInfo to minimize confusion.
This is the lint-breaking protobuf API change in this diff.

Next diffs:
1. Implement query result forwarder
2. Hook query result forwarder into ExecuteScript
3. Change compiler to produce plans whose final sinks are GRPC sinks that write to the qb, rather than memory sinks.
4. Kelvin should not use ReceiveAgentQueryResult endpoint when all sinks are GRPC sinks
5. Deprecate ReceiveAgentQueryResult and stale code in ExecuteScript

Test Plan: added unit tests

Reviewers: philkuz, zasgar, michelle, #engineering

Reviewed By: zasgar, #engineering

JIRA Issues: PP-2117

Differential Revision: https://phab.corp.pixielabs.ai/D5944

GitOrigin-RevId: 2cbb7aaa5961439b18eab7bd31cf601c07d84398

Summary:
Some more changes to TransferResultChunk message, that were discovered as I started to implement the query broker side of things.
1. grpc_source_id and table_name only apply to the case where TransferResultChunk is sending a row batch, so update the message structure so that they only occur alongside a row batch.
2. Add a field for schema, since there was no field for result schema before.
3. The naming of ExecutionAndTimingInfo message alongside the execution_and_timing_info field was producing a confusing collision of names in the resulting go proto. I renamed the message to QueryExecutionAndTimingInfo so that wouldn't occur anymore.

Test Plan: updated unit tests

Reviewers: philkuz, michelle, zasgar, #engineering

Reviewed By: zasgar, #engineering

JIRA Issues: PP-2117

Differential Revision: https://phab.corp.pixielabs.ai/D5952

GitOrigin-RevId: b3e5c09d9ca57b80069eead05a6f6fef3c3489a5

Summary: Required for C++ tracing.

Test Plan: Tests modified.

Reviewers: yzhao, #engineering

Reviewed By: yzhao, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5946

GitOrigin-RevId: 629dec9d256754232394cd3bf24426c3581287ae

Summary:
This is used to determine whether or not to index overloaded functions, which reads parameter types
in addition to function names (diffs are coming).

Test Plan: Jenkins

Reviewers: oazizi, #engineering

Reviewed By: oazizi, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5937

GitOrigin-RevId: 83db183bd72c36d75b452e96ce485561a012d715

Summary: Added proto message to serialize documentation of UDFs from C++.

Test Plan: N/A

Reviewers: philkuz, #engineering

Reviewed By: philkuz, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5947

GitOrigin-RevId: 0e5aa24910d38af1d18a9e1bb0d9ede5760bd91b

Summary:
The prototool excludes needed updating after files were moved.

Now just exclude the physicalpb, since that one is a private protobuf.
Rest are now public, so enforce them.

Test Plan: None

Reviewers: michelle, yzhao, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5953

GitOrigin-RevId: 9a3f32ab7c35756340cea09a6393b5cf55fc6bbf

Summary: Organize related tests, and refactor repetitive code into functions.

Test Plan: Jenkins

Reviewers: yzhao, #engineering

Reviewed By: yzhao, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5933

GitOrigin-RevId: de4276999e1f0214e0f1c2e9656385519fa685ae

Summary: This diff adds support for tracing struct types in C/C++.

Test Plan: End-to-end test added in stirling_bpf_test

Reviewers: #engineering, yzhao

Reviewed By: #engineering, yzhao

Differential Revision: https://phab.corp.pixielabs.ai/D5929

GitOrigin-RevId: f2dee17f2a9c2f1b7d431768e1af54606560883a

Summary:
A function is considered utility if it's entirely operate on DWARF concept, and does not consider
any language-specific, or things relevant to our intention.

Test Plan: Jenkins

Reviewers: oazizi, #engineering

Reviewed By: oazizi, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5935

GitOrigin-RevId: 66a33a2cf7fd908275c3044ef5ec50fe7b40b374

Summary:
when the user selects an entity, we automatically create a new, empty tabstop that we move their cursor to.
however, this tabstop is not populated with any suggestions, so the interaction seems a bit weird until the user starts typing again.

updated the backend so that if we do create new, empty tabstop, we also make a call to elastic to get a set of suggestions.

Test Plan: unit test, ran in cloud dev and checked that it did the right thing

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5941

GitOrigin-RevId: 5233d830c8a98362e3ef4da478220bd945a15a5b

Summary:
until we have the script indexer up and ready, we've been matching scripts using the sahilm fuzzy search golang module.
this fuzzy search doesn't match any words when given an empty input string. in our case, we want it to match all scripts. these scripts will be filtered down further depending on the type of the input arguments already specified.
i updated the suggester code so that this is the case.
again, all of this script matching is just temporary until we have the script indexer... which probably won't be for a while

Test Plan: tried in cloud dev

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5945

GitOrigin-RevId: 7a7f8fecf4f5bbe0a8cb78ea8bd40c0da6cb2364

Summary:
We should show some indication for whether the user has created a valid/invalid command.
to follow what we do in the CLI, I added a green/red border.
another option is to color the "px" icon in the command box, for something more subtle.

Test Plan:
screenshots:
{F91513}

{F91514}

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5942

GitOrigin-RevId: 1e71e2a0ba82d49a09439374bd8ef7769a437325

Summary: Will be used for testing changes to the query broker, specifically the implementation of the API TransferResultChunk.

Test Plan: n/a

Reviewers: michelle, zasgar, philkuz, #engineering

Reviewed By: philkuz, #engineering

JIRA Issues: PP-2117

Differential Revision: https://phab.corp.pixielabs.ai/D5940

GitOrigin-RevId: 73a7c91abf5e17314db948014b43f55c4bbf2a72

Summary:
proto serializer for blocking agg relies on relations being right, that wasn't the case here. Hack that fixese it for now but need clarification on usage before providing something better.

Also fixed the way we load aggregate expressions so that we load the functions correctly instead of trying to evaluate the children.

Test Plan: passes previously broken test and also fails in execution rather than in compilation for partial aggregates (which is expected, we want to update the aggregation operator).

Reviewers: zasgar, #engineering, nserrino

Reviewed By: #engineering, nserrino

Subscribers: nserrino

Differential Revision: https://phab.corp.pixielabs.ai/D5921

GitOrigin-RevId: 1555dc1375116203070ddc4d23201741b9d1006f

Summary:
when the command input loads up for the first time, we should make an API call to the backend to get a list of initial suggestions for the empty input.
there is also some cleanup for the debouncing

Test Plan: tried it in webpack

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5939

GitOrigin-RevId: 26572009d34f93030de623cf4dfee41aa1f22076

Summary: spaces weren't showing up properly in some cases in the autocomplete UI. this is because the span containing the empty space was rendering with a width of 0. instead, if we use a non-breaking space, the span should render with the proper width

Test Plan: ran webpack

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5938

GitOrigin-RevId: 3e2dcdc3de03f5b90deea9223426308f232d4c89

Summary:
we want to update the UDATester to automatically test for a few more things:
- serialization/deserialization. when Expect() is called, we should test that deserializing a serialized value produces the correct result.
- Merging, in a random order to test that our operations are commutative + associative.

Test Plan: ran math_ops test

Reviewers: zasgar, philkuz, nserrino, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5932

GitOrigin-RevId: 6f6820236b70980116d0f5d705a5929801ab39f6

Summary:
we recently updated the build so that it pushes the images to a public repo for official releases...
unfortunately, there was a hard-coded path for the updater image hidden in our files.
this broke 0.4.0, but i put out a temporary fix by uploading the correct image to the pl-dev-infra registry for now.

this diff takes the image path from the cloudconnector image and uses that image path for the updater image.

Test Plan: ran skaffold

Reviewers: zasgar, #engineering

Reviewed By: zasgar, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D5926

GitOrigin-RevId: 23b3641dec96ec69440f388c14729a9a314c0e6c

Summary:
Depends on D5927. This protobuf api breaking change should also still be ok because these services are only used internally in Vizier and we don't yet have rolling upgrades.
KelvinService isn't accurate anymore, it is more accurate to call it ResultSinkService because it is the api for any service that implements the apis necessary to receive a result chunk (aka row batch or exec stats). This group will include both Kelvin and Query broker in the near term.

Test Plan: existing

Reviewers: zasgar, michelle, philkuz, #engineering

Reviewed By: philkuz, #engineering

JIRA Issues: PP-2117

Differential Revision: https://phab.corp.pixielabs.ai/D5928

GitOrigin-RevId: dcfa43436f928daea8c9e0620a9ec9c0f056e3d5

Summary:
TransferResultChunk is the new API for streaming row batches to their recipients. It supports sending query stats as well as sending result data to non-Carnot sources, such as the query broker. This change will allow GRPCSinks to send directly to the query broker, which will support streaming mode and also remove the need for Kelvins to buffer up the data before sending it to the query broker (even in batch mode, this will be good because it will allow results to show up as they are available).

Changed the field names on some protobufs so that the terminology and naming would be more consistent and less confusing going forward. This should be an ok set of changes because it is all internal to Vizier.

Test Plan: existing tests

Reviewers: zasgar, michelle, philkuz, #engineering

Reviewed By: philkuz, #engineering

JIRA Issues: PP-2117

Differential Revision: https://phab.corp.pixielabs.ai/D5927

GitOrigin-RevId: acf29b492312a0f9780453b58d3e2b661011513d