standard_health 0.7.0 → 0.7.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: 4b9de071226844b0a4c1304b7b81d21ba11b963e3c4ad082aebd05a82a24fa92
4
- data.tar.gz: dd4bab92dade63959f66db19ee8467c753e39a1f77dd12706adff721460bc14e
3
+ metadata.gz: 162374f071e6634d43876ef75f8fa6f50b1ec7bfd6a8d01bd94a91c854151d09
4
+ data.tar.gz: 58c40844359768a3d905c75575599dc1bf61dab15ebabc372c3e5e56cd4e169d
5
5
  SHA512:
6
- metadata.gz: 1b6431be6025cdfb36f4059d1d295464c43d0daf927155cd53953c34cbac683ce04fbd810acc7217765be574fa4ea85eb59a8186d9326c617225705efdf878ef
7
- data.tar.gz: 623ea6e63b5f4e11c281ad64d08e60ec10e858b960adcc6758ffa98970cc5e92cd4c0e64406d679aad2d44e387800de88f2bef521f5513899188ae81e811cba9
6
+ metadata.gz: a93890d618342991d02699f9a68d0cf432596f4fba5eff7659d4d4c8675d22e5e0bac3d996abed41e02ddd7f53b84c771f21e093a6c6e22b39404ec9ccc7498c
7
+ data.tar.gz: 13b347704251b47c6a877ce5fb559686d9fee639610109300da1fd989627010a3823f678ab420330a26ac851548647da3ab12405081cd13a2d9024c90d4bb5dc
data/CHANGELOG.md CHANGED
@@ -7,6 +7,47 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
7
7
 
8
8
  ## [Unreleased]
9
9
 
10
+ ## [0.7.1] - 2026-09-25
11
+
12
+ ### Fixed
13
+
14
+ - **Skipped checks no longer degrade the aggregate.** A check that returns
15
+ `status: :skipped` itself means "not applicable here" (the feature it
16
+ covers is not configured or not enforced). It is now **neutral** in the
17
+ roll-up on every tier (`/ready` and the aggregate `/health`): it neither
18
+ degrades nor fails the status. Before, it counted like a failing
19
+ non-critical check, which held sidekick-web's `/health` at `degraded`
20
+ permanently while attestation was unconfigured. The row still renders in
21
+ `checks[]` with `"status": "skipped"`, and `check.completed` still fires,
22
+ so it stays visible. It is no longer listed in the `failed[]` / `failures`
23
+ of `ready.evaluated` / `aggregate.evaluated`, so the Logger and Sentry
24
+ notifiers don't name it as failing when something else degrades.
25
+ - **A critical check that reports `:skipped` is neutral too**, on purpose.
26
+ "Not applicable" says nothing about whether the instance can serve, so it
27
+ must not pull it out of rotation. Return `:fail` for a state that should
28
+ page.
29
+ - Real failures roll up exactly as before: a failing critical check is
30
+ `unavailable` (503), a failing non-critical one `degraded`, including
31
+ alongside a skipped check.
32
+ - **Budget skips are unchanged.** A check the `total_check_budget` never
33
+ reached was not performed, which is not the same as healthy. It still
34
+ floors the roll-up at `degraded` (never `unavailable`) and is still listed
35
+ in `failed[]`. Its row now carries `budget_exhausted: true`, which is how
36
+ the gem tells the two kinds of skip apart.
37
+ - `/diagnostics/env` is unaffected. Its roll-up is over env-audit rows and
38
+ assertions (`:ok` / `:warn` / `:error`), which have no skipped state.
39
+
40
+ ### Changed
41
+
42
+ - **Requires Rails 8.1** (`rails >= 8.1`, was `>= 8.0`). Every consumer app
43
+ runs 8.1; 8.0 was never exercised in CI. Ruby 3.4 remains supported.
44
+
45
+ ### Upgrade notes (0.7.0 → 0.7.1)
46
+
47
+ - **No required host change.** An app whose aggregate was `degraded` only
48
+ because of a self-reported skip now reports `ok`. Monitors or smoke tests
49
+ that expect `degraded` in that state need updating (sidekick-web).
50
+
10
51
  ## [0.7.0] - 2026-09-24
11
52
 
12
53
  The Phase 4 release. 0.6 deprecated nothing, so nothing is removed. It adds
data/README.md CHANGED
@@ -106,7 +106,7 @@ end
106
106
  | Status | HTTP | When |
107
107
  |---|---|---|
108
108
  | `unavailable` | 503 | a critical check failed, or a `:critical` circuit is red |
109
- | `degraded` | 200 | a non-critical check failed or was skipped, or the circuit roll-up is `:degraded` |
109
+ | `degraded` | 200 | a non-critical check failed, a check was skipped by the total budget, or the circuit roll-up is `:degraded` |
110
110
  | `ok` | 200 | otherwise |
111
111
 
112
112
  **The aggregate tier re-runs your readiness checks by default.** With
@@ -806,9 +806,25 @@ go in `extra:`.
806
806
 
807
807
  The orchestrator should pull the instance out of rotation only on 503; degraded means "still serving, page someone."
808
808
 
809
- `skipped` also exists, for a check the total budget never reached (see
810
- Timeouts). A skip floors the roll-up at `degraded` and **never** produces
811
- `unavailable`, even for a critical check.
809
+ `skipped` also exists, in two flavours that roll up differently:
810
+
811
+ - **A check that returns `status: :skipped` itself** — "not applicable here"
812
+ (the feature it covers isn't configured or enforced). This is **neutral**
813
+ (since 0.7.1): it neither degrades nor fails the roll-up, whether the check
814
+ is critical or not, and it is left out of the evaluation events' `failed[]`.
815
+ It still renders in `checks[]` with `"status": "skipped"` and still emits
816
+ `check.completed`, so it stays visible. A critical check is neutral too on
817
+ purpose: "not applicable" says nothing about whether the instance can serve,
818
+ so it must not pull it out of rotation — and it must not mark it healthy
819
+ *because of* that check either; it simply doesn't count. If the state should
820
+ page, return `:fail`.
821
+ - **A check the total budget never reached** (see Timeouts) — rendered with
822
+ `"budget_exhausted": true`. That check was not performed, which is not the
823
+ same as healthy, so it floors the roll-up at `degraded` and **never**
824
+ produces `unavailable`, even for a critical check.
825
+
826
+ A real failure still rolls up exactly as before alongside any skip: a failing
827
+ critical check is `unavailable`, a failing non-critical one `degraded`.
812
828
 
813
829
  ## Failure detail is redacted
814
830
 
@@ -934,8 +950,8 @@ slow-but-fine starts reporting `:fail`, and for a critical check that pulls
934
950
  the instance out of rotation. Pick values from observed `latency_ms` rather
935
951
  than intuition; that is what the `check.completed` events are for.
936
952
 
937
- Checks the total budget never reaches report `:skipped`, never silently `:ok`.
938
- A skip alone floors the roll-up at `degraded` — otherwise a slow *non-critical*
953
+ Checks the total budget never reaches report `:skipped` (with
954
+ `budget_exhausted: true`), never silently `:ok`. A budget skip alone floors the roll-up at `degraded` — otherwise a slow *non-critical*
939
955
  check could exhaust the budget, leave the database check unrun, and pull a
940
956
  healthy instance out of rotation.
941
957
 
@@ -18,8 +18,10 @@ module StandardHealth
18
18
  #
19
19
  # Status roll-up, in the estate's readiness vocabulary:
20
20
  # :unavailable (503) — a critical check failed, or a :critical circuit is RED
21
- # :degraded (200) — a non-critical check failed / was skipped, or the
22
- # circuit roll-up is :degraded
21
+ # :degraded (200) — a non-critical check failed, a check was skipped
22
+ # by the total budget, or the circuit roll-up is
23
+ # :degraded. A check that reports :skipped itself
24
+ # (not applicable) is neutral — see Aggregator.
23
25
  # :ok (200) — otherwise
24
26
  #
25
27
  # StandardCircuit's own word for its worst state is `:critical`; it is
@@ -92,7 +94,7 @@ module StandardHealth
92
94
  # notifier logs this one (so redacted messages still reach logs); Sentry
93
95
  # and Metrics ignore it.
94
96
  def emit_evaluation(report, rows, circuits, started)
95
- failing = rows.reject { |r| r[:status] == :ok }
97
+ failing = Aggregator.failing_rows(rows)
96
98
  payload = {
97
99
  status: report[:status],
98
100
  duration_ms: ((Process.clock_gettime(Process::CLOCK_MONOTONIC) - started) * 1000).round,
@@ -14,11 +14,23 @@ module StandardHealth
14
14
  # Runs all registered checks and rolls them up into a single status.
15
15
  #
16
16
  # Status semantics:
17
- # :ok — every check returned :ok
17
+ # :ok — every check returned :ok, or reported :skipped itself
18
18
  # :degraded — at least one non-critical check failed, OR a check was
19
19
  # skipped because the total budget ran out
20
20
  # :unavailable — at least one critical check failed
21
21
  #
22
+ # Two kinds of :skipped row exist, and they roll up differently (0.7.1):
23
+ #
24
+ # * A check that RETURNS `status: :skipped` is saying "not applicable
25
+ # here" (the feature it covers is not configured or not enforced). That
26
+ # is NEUTRAL — it neither degrades nor fails the tier, whatever its
27
+ # `critical:` flag, and it is not listed in an evaluation event's
28
+ # `failed[]`. The row still renders with status "skipped", and
29
+ # `check.completed` still fires, so it stays visible.
30
+ # * A check the total budget never REACHED (`budget_exhausted: true` on
31
+ # the row) was not performed at all, and an unperformed check is not a
32
+ # healthy one. It floors the tier at :degraded — never :unavailable.
33
+ #
22
34
  # The aggregator never raises. Each check is invoked through `safe_run`
23
35
  # which catches `StandardError` so a buggy custom check cannot take down
24
36
  # /ready. Instrumentation is held to the same bar — every emit is wrapped
@@ -55,7 +67,7 @@ module StandardHealth
55
67
  duration_ms = ((monotonic - started) * 1000).round
56
68
  status = overall_status(check_rows)
57
69
 
58
- failing = check_rows.reject { |r| r[:status] == :ok }
70
+ failing = failing_rows(check_rows)
59
71
 
60
72
  return { status: status, checks: check_rows, generated_at: now.iso8601 } unless tier == :ready
61
73
 
@@ -81,6 +93,21 @@ module StandardHealth
81
93
  }
82
94
  end
83
95
 
96
+ # The rows an evaluation event reports as failing: everything that is not
97
+ # :ok, except a check that reported :skipped itself (not applicable, so
98
+ # not a failure). A budget skip IS listed — it is the only place a skip
99
+ # shows up in the transition-gated event. @api private (also used by
100
+ # AggregateReport).
101
+ def self.failing_rows(rows)
102
+ rows.reject { |r| r[:status] == :ok || neutral_skip?(r) }
103
+ end
104
+
105
+ # A row the check itself reported as :skipped — "not applicable here".
106
+ # Neutral in the roll-up. A budget skip is not neutral. @api private
107
+ def self.neutral_skip?(row)
108
+ row[:status] == :skipped && !row[:budget_exhausted]
109
+ end
110
+
84
111
  # Per-failure detail for an evaluation event. Public (but @api private) so
85
112
  # the aggregate tier's evaluation event carries the same shape.
86
113
  # @api private
@@ -177,12 +204,15 @@ module StandardHealth
177
204
  private_class_method :budget_exhausted?
178
205
 
179
206
  # A check the total budget never reached. `:skipped`, never silently
180
- # `:ok` — an unperformed check is not a healthy one.
207
+ # `:ok` — an unperformed check is not a healthy one. `budget_exhausted`
208
+ # is what separates it from a check that reported :skipped itself (which
209
+ # is neutral); see overall_status.
181
210
  def self.skipped_row(reg, budget)
182
211
  {
183
212
  name: reg.name,
184
213
  critical: reg.critical,
185
214
  status: :skipped,
215
+ budget_exhausted: true,
186
216
  error: "skipped — total check budget of #{budget}s exhausted"
187
217
  }
188
218
  end
@@ -191,10 +221,16 @@ module StandardHealth
191
221
  def self.overall_status(rows)
192
222
  return :ok if rows.empty?
193
223
 
194
- failures = rows.reject { |r| r[:status] == :ok }
224
+ # A check that reported :skipped itself is not applicable here, so it is
225
+ # NEUTRAL: it must not degrade the tier (sidekick-web's non-critical
226
+ # attestation check held /health at "degraded" forever while attestation
227
+ # was simply not configured), and a critical one must not fail it
228
+ # either — "not applicable" says nothing about whether the app can
229
+ # serve. It stays in checks[] as "skipped" for visibility.
230
+ failures = failing_rows(rows)
195
231
  return :ok if failures.empty?
196
232
 
197
- # A SKIP MUST NEVER PRODUCE :unavailable, even for a critical check.
233
+ # A BUDGET SKIP MUST NEVER PRODUCE :unavailable, even for a critical check.
198
234
  # Otherwise a slow *non-critical* check could exhaust the budget, leave
199
235
  # the database check unrun, and pull a perfectly healthy instance out of
200
236
  # rotation — a self-inflicted outage caused by the safety mechanism.
@@ -11,6 +11,16 @@ module StandardHealth
11
11
  #
12
12
  # { status: :fail, error: "connection refused" }
13
13
  #
14
+ # or, when the check does not apply to this deployment (the feature it
15
+ # covers is not configured or not enforced):
16
+ #
17
+ # { status: :skipped }
18
+ #
19
+ # A self-reported :skipped is NEUTRAL in the roll-up (0.7.1): it never
20
+ # degrades or fails the tier, even for a critical check, but it still
21
+ # renders in `checks[]` as "skipped". Return :fail, not :skipped, for
22
+ # anything that should page.
23
+ #
14
24
  # The `with_timing` helper wraps a block, captures latency, and converts
15
25
  # any unhandled `StandardError` into a `:fail` row so subclasses don't
16
26
  # have to repeat the pattern.
@@ -192,7 +192,9 @@ module StandardHealth
192
192
  attr_accessor :default_check_timeout
193
193
 
194
194
  # Budget across all checks, evaluated BEFORE each check starts. Checks not
195
- # reached are reported :skipped, and a skip alone floors the roll-up at
195
+ # reached are reported :skipped with `budget_exhausted: true` (unlike a
196
+ # check that reports :skipped itself, which is neutral), and such a skip
197
+ # alone floors the roll-up at
196
198
  # :degraded — never :unavailable. Otherwise a slow NON-critical check could
197
199
  # exhaust the budget, leave a critical check unrun, and pull a healthy app
198
200
  # out of rotation. nil = no budget.
@@ -1,5 +1,5 @@
1
1
  # frozen_string_literal: true
2
2
 
3
3
  module StandardHealth
4
- VERSION = "0.7.0"
4
+ VERSION = "0.7.1"
5
5
  end
metadata CHANGED
@@ -1,7 +1,7 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: standard_health
3
3
  version: !ruby/object:Gem::Version
4
- version: 0.7.0
4
+ version: 0.7.1
5
5
  platform: ruby
6
6
  authors:
7
7
  - Jaryl Sim
@@ -15,14 +15,14 @@ dependencies:
15
15
  requirements:
16
16
  - - ">="
17
17
  - !ruby/object:Gem::Version
18
- version: '8.0'
18
+ version: '8.1'
19
19
  type: :runtime
20
20
  prerelease: false
21
21
  version_requirements: !ruby/object:Gem::Requirement
22
22
  requirements:
23
23
  - - ">="
24
24
  - !ruby/object:Gem::Version
25
- version: '8.0'
25
+ version: '8.1'
26
26
  description: StandardHealth is a mountable Rails engine providing /alive, /ready,
27
27
  and /diagnostics/env endpoints, with a configuration block for registering custom
28
28
  checks and a DSL for declaring required and recommended environment variables.
@@ -88,7 +88,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
88
88
  - !ruby/object:Gem::Version
89
89
  version: '0'
90
90
  requirements: []
91
- rubygems_version: 4.0.3
91
+ rubygems_version: 4.0.10
92
92
  specification_version: 4
93
93
  summary: A drop-in health check and environment-spec engine for Rails 8 host apps.
94
94
  test_files: []