fix(codecs): native JSON serialization/deserialization for special f64 values

[pront]

closes: #18558

Problem

f64::INFINITY, f64:: NEG_INFINITY and f64::NAN are all serialized to JSON null. The original value cannot be recovered at deserialization.

Also, our generated fixtures exclude special f64 values so this issue didn't surface earlier. Created #18666.

Solution

Use serde's https://serde.rs/field-attrs.html#serialize_with for serializing/deserializing upper_limmit.

Alternatives considered

• Change the upper limit type to Option<f64>.
• Use SerializeDisplay and DeserializeFromStr macros.
• Do not use f64::INFINITY and other special values. Replace with a large constant.

All of the above alternatives (further) break backwards compatibility.

[netlify]

✅ Deploy Preview for vrl-playground ready!

Name	Link
🔨 Latest commit	3421949
🔍 Latest deploy log	https://app.netlify.com/sites/vrl-playground/deploys/651dccfe2e48b9000825a198
😎 Deploy Preview	https://deploy-preview-18650--vrl-playground.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[netlify]

✅ Deploy Preview for vector-project canceled.

Name	Link
🔨 Latest commit	3421949
🔍 Latest deploy log	https://app.netlify.com/sites/vector-project/deploys/651dccfec512c000081dcce0

[aws-amplify-us-east-1]

This pull request is automatically being deployed by Amplify Hosting (learn more).

Access this pull request here: https://pr-18650.d1hoyoksbulg25.amplifyapp.com

[aws-amplify-us-east-1]

This pull request is automatically being deployed by Amplify Hosting (learn more).

Access this pull request here: https://pr-18650.d2lr4eds605rpz.amplifyapp.com

[aws-amplify-us-east-1]

This pull request is automatically being deployed by Amplify Hosting (learn more).

Access this pull request here: https://pr-18650.d1a7j77663uxsc.amplifyapp.com

[bruceg]

This looks reasonable, with one caveat below.

[datadog-vectordotdev]

Datadog Report

Branch report: pront/fix-bucket-ser-der
Commit report: 9665f53

✅ vector: 0 Failed, 0 New Flaky, 1981 Passed, 0 Skipped, 1m 18.71s Wall Time

[bruceg]

This looks good and functional. I left a suggestion for avoiding allocating and freeing a temporary string in these code paths, but that isn't a correctness issue.

[github-actions]

Regression Detector Results

Run ID: f0086fe0-64e5-4382-ae0b-68b8633a8634
Baseline: 1452d54
Comparison: 39b9298
Total vector CPUs: 7

Explanation

A regression test is an integrated performance test for vector in a repeatable rig, with varying configuration for vector. What follows is a statistical summary of a brief vector run for each configuration across SHAs given above. The goal of these tests are to determine quickly if vector performance is changed and to what degree by a pull request.

Because a target's optimization goal performance in each experiment will vary somewhat each time it is run, we can only estimate mean differences in optimization goal relative to the baseline target. We express these differences as a percentage change relative to the baseline target, denoted "Δ mean %". These estimates are made to a precision that balances accuracy and cost control. We represent this precision as a 90.00% confidence interval denoted "Δ mean % CI": there is a 90.00% chance that the true value of "Δ mean %" is in that interval.

We decide whether a change in performance is a "regression" -- a change worth investigating further -- if both of the following two criteria are true:

1. The estimated |Δ mean %| ≥ 5.00%. This criterion intends to answer the question "Does the estimated change in mean optimization goal performance have a meaningful impact on your customers?". We assume that when |Δ mean %| < 5.00%, the impact on your customers is not meaningful. We also assume that a performance change in optimization goal is worth investigating whether it is an increase or decrease, so long as the magnitude of the change is sufficiently large.

2. Zero is not in the 90.00% confidence interval "Δ mean % CI" about "Δ mean %". This statement is equivalent to saying that there is at least a 90.00% chance that the mean difference in optimization goal is not zero. This criterion intends to answer the question, "Is there a statistically significant difference in mean optimization goal performance?". It also means there is no more than a 10.00% chance this criterion reports a statistically significant difference when the true difference in mean optimization goal is zero -- a "false positive". We assume you are willing to accept a 10.00% chance of inaccurately detecting a change in performance when no true difference exists.

The table below, if present, lists those experiments that have experienced a statistically significant change in mean optimization goal performance between baseline and comparison SHAs with 90.00% confidence OR have been detected as newly erratic. Negative values of "Δ mean %" mean that baseline is faster, whereas positive values of "Δ mean %" mean that comparison is faster. Results that do not exhibit more than a ±5.00% change in their mean optimization goal are discarded. An experiment is erratic if its coefficient of variation is greater than 0.1. The abbreviated table will be omitted if no interesting change is observed.

Changes in experiment optimization goals with confidence ≥ 90.00% and |Δ mean %| ≥ 5.00%:

experiment	goal	Δ mean %	confidence
otlp_http_to_blackhole	ingress throughput	+6.23	100.00%

Fine details of change detection per experiment.

experiment	goal	Δ mean %	Δ mean % CI	confidence
otlp_http_to_blackhole	ingress throughput	+6.23	[+6.06, +6.39]	100.00%
syslog_log2metric_splunk_hec_metrics	ingress throughput	+3.28	[+3.12, +3.44]	100.00%
socket_to_socket_blackhole	ingress throughput	+2.87	[+2.79, +2.95]	100.00%
syslog_regex_logs2metric_ddmetrics	ingress throughput	+2.03	[+1.62, +2.43]	100.00%
otlp_grpc_to_blackhole	ingress throughput	+1.80	[+1.68, +1.92]	100.00%
datadog_agent_remap_blackhole_acks	ingress throughput	+1.72	[+1.58, +1.85]	100.00%
datadog_agent_remap_datadog_logs	ingress throughput	+1.49	[+1.33, +1.65]	100.00%
syslog_splunk_hec_logs	ingress throughput	+1.45	[+1.37, +1.53]	100.00%
datadog_agent_remap_blackhole	ingress throughput	+1.18	[+1.07, +1.30]	100.00%
http_to_http_acks	ingress throughput	+1.09	[+0.93, +1.25]	100.00%
syslog_humio_logs	ingress throughput	+1.09	[+0.95, +1.23]	100.00%
splunk_hec_route_s3	ingress throughput	+0.95	[+0.42, +1.48]	99.67%
syslog_loki	ingress throughput	+0.75	[+0.64, +0.85]	100.00%
datadog_agent_remap_datadog_logs_acks	ingress throughput	+0.36	[+0.20, +0.51]	99.98%
splunk_hec_to_splunk_hec_logs_acks	ingress throughput	+0.03	[-0.13, +0.20]	26.17%
splunk_hec_to_splunk_hec_logs_noack	ingress throughput	+0.02	[-0.12, +0.16]	18.23%
file_to_blackhole	egress throughput	+0.01	[-1.07, +1.10]	1.70%
http_to_http_json	ingress throughput	-0.00	[-0.05, +0.05]	3.86%
splunk_hec_indexer_ack_blackhole	ingress throughput	-0.00	[-0.17, +0.16]	2.22%
syslog_log2metric_humio_metrics	ingress throughput	-0.39	[-0.55, -0.22]	99.99%
fluent_elasticsearch	ingress throughput	-0.50	[-0.80, -0.20]	99.38%
http_to_s3	ingress throughput	-0.71	[-0.99, -0.42]	100.00%
http_to_http_noack	ingress throughput	-1.33	[-1.52, -1.15]	100.00%
enterprise_http_to_http	ingress throughput	-1.35	[-1.51, -1.18]	100.00%
http_text_to_http_json	ingress throughput	-1.53	[-1.66, -1.41]	100.00%

[folliehiyuki]

Hi, @pront!

I'm trying to build the latest release 0.34.1 on AlpineLinux, and I'm hitting a test failure in serialize_aggregated_histogram test case introduced by this PR. Can you give me some hints on how to fix it?

The package update attempt is here: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/56438.

And this is the full build+test log from CI: https://gitlab.alpinelinux.org/FollieHiyuki/aports/-/jobs/1199494

A snippet of the failure part of the log:

...
------------
     Summary [  17.948s] 239/738 tests run: 238 passed, 1 failed, 4 skipped
  TRY 4 FAIL [   0.005s] codecs encoding::format::native_json::tests::serialize_aggregated_histogram
--- TRY 4 STDOUT:        codecs encoding::format::native_json::tests::serialize_aggregated_histogram ---
running 1 test
test encoding::format::native_json::tests::serialize_aggregated_histogram ... FAILED
failures:
failures:
    encoding::format::native_json::tests::serialize_aggregated_histogram
test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 94 filtered out; finished in 0.00s
--- TRY 4 STDERR:        codecs encoding::format::native_json::tests::serialize_aggregated_histogram ---
thread 'encoding::format::native_json::tests::serialize_aggregated_histogram' panicked at 'assertion failed: `(left == right)`
  left: `b"{\"metric\":{\"name\":\"histogram\",\"kind\":\"absolute\",\"aggregated_histogram\":{\"buckets\":[{\"upper_limit\":\"-inf\",\"count\":0},{\"upper_limit\":2.0,\"count\":1},{\"upper_limit\":\"inf\",\"count\":0}],\"count\":1,\"sum\":1.0}}}"`,
 right: `"{\"metric\":{\"aggregated_histogram\":{\"buckets\":[{\"count\":0,\"upper_limit\":\"-inf\"},{\"count\":1,\"upper_limit\":2.0},{\"count\":0,\"upper_limit\":\"inf\"}],\"count\":1,\"sum\":1.0},\"kind\":\"absolute\",\"name\":\"histogram\"}}"`', lib/codecs/src/encoding/format/native_json.rs:107:9

[pront]

Hi, @pront!

I'm trying to build the latest release 0.34.1 on AlpineLinux, and I'm hitting a test failure in serialize_aggregated_histogram test case introduced by this PR. Can you give me some hints on how to fix it?

The package update attempt is here: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/56438.

And this is the full build+test log from CI: https://gitlab.alpinelinux.org/FollieHiyuki/aports/-/jobs/1199494

A snippet of the failure part of the log:

...
------------
     Summary [  17.948s] 239/738 tests run: 238 passed, 1 failed, 4 skipped
  TRY 4 FAIL [   0.005s] codecs encoding::format::native_json::tests::serialize_aggregated_histogram
--- TRY 4 STDOUT:        codecs encoding::format::native_json::tests::serialize_aggregated_histogram ---
running 1 test
test encoding::format::native_json::tests::serialize_aggregated_histogram ... FAILED
failures:
failures:
    encoding::format::native_json::tests::serialize_aggregated_histogram
test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 94 filtered out; finished in 0.00s
--- TRY 4 STDERR:        codecs encoding::format::native_json::tests::serialize_aggregated_histogram ---
thread 'encoding::format::native_json::tests::serialize_aggregated_histogram' panicked at 'assertion failed: `(left == right)`
  left: `b"{\"metric\":{\"name\":\"histogram\",\"kind\":\"absolute\",\"aggregated_histogram\":{\"buckets\":[{\"upper_limit\":\"-inf\",\"count\":0},{\"upper_limit\":2.0,\"count\":1},{\"upper_limit\":\"inf\",\"count\":0}],\"count\":1,\"sum\":1.0}}}"`,
 right: `"{\"metric\":{\"aggregated_histogram\":{\"buckets\":[{\"count\":0,\"upper_limit\":\"-inf\"},{\"count\":1,\"upper_limit\":2.0},{\"count\":0,\"upper_limit\":\"inf\"}],\"count\":1,\"sum\":1.0},\"kind\":\"absolute\",\"name\":\"histogram\"}}"`', lib/codecs/src/encoding/format/native_json.rs:107:9

Hi @folliehiyuki, this looks like a JSON key/value ordering issue. However the data look correct. We could relax the final assertion on the test. I might have time for this later this week. Also, PRs are always welcome.