coffey-labs/cairnobs
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Specify what aggregate_count emits

Browse Source 
The last unanswered action, and the only one whose output is not the
input with edits -- it emits a record that never existed, which is why
it was deferred twice.

It emits the window's first record unchanged, tagged with
cairnobs.aggregated, cairnobs.count, and the observed window bounds.
That follows the convention the agent already uses for heartbeat and
host-metrics records rather than inventing a second synthetic-record
mechanism, and keeping the first record intact means a reader sees a
real example of what was collapsed instead of an invented summary.

It tags even when the count is one. Emitting a bare record there would
be tidier and would make cairnobs.count present only sometimes, so
summing it silently breaks on quiet windows. window_last is the last
record that actually contributed, never window_start + window_ms,
because a window flushed early must not claim an end that never
happened.

Specifying it surfaced a problem the other nine actions do not have.
Windows are measured on record time, so a window can only be closed by a
later record arriving. suppress_duplicates never has anything pending;
aggregate_count holds state, so a matching stream that goes quiet leaves
its aggregate unemitted indefinitely -- data loss dressed as latency.
Emission therefore has a second trigger, end of stream, which the corpus
defines as an implicit flush after the last input and which production
gets from the batch flush. The cost is stated rather than hidden:
window_ms becomes a maximum, not a guarantee, and one burst can produce
more than one aggregate.

And it has a consequence nobody should meet in production first: stats
count undercounts aggregated data silently, so every panel and alert
counting rows changes meaning the moment a rule aggregates the data
behind it. Nothing here fixes that. The correct idiom is summing
cairnobs.count; teaching the query layer to do it automatically is a
Phase 2 change to the IR, recorded as the open question this decision
leaves in its place rather than quietly inherited.

Six cases added, corpus at 45. The validator's unspecified-action guard
stays in place with an empty set, still rejecting anything added to it.

Signed-off-by: John Coffey <[email protected]>
This commit is contained in:
jcoffey-dev
2026-09-04 21:58:40 -07:00
parent 17256634db
commit ec4b860ba8
9 changed files with 494 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files
processing/conformance/cases/aggregate_count_closes_a_window_on_a_later_record.json
+69
View File
@@ -0,0 +1,69 @@
{
"name": "aggregate_count_closes_a_window_on_a_later_record",
"description": "A record at or past window_start + window_ms flushes the pending aggregate first, then opens the next window. The second window is flushed at end of stream.",
"rules": [
{
"match": [],
"actions": [
{
"action": "aggregate_count",
"window_ms": 5000
}
]
}
],
"inputs": [
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "x",
"attributes": {}
},
{
"timestamp_unix_nano": 1000000000,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "x",
"attributes": {}
},
{
"timestamp_unix_nano": 6000000000,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "x",
"attributes": {}
}
],
"expect": [
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "x",
"attributes": {
"cairnobs.aggregated": "true",
"cairnobs.count": "2",
"cairnobs.window_start_unix_nano": "0",
"cairnobs.window_last_unix_nano": "1000000000"
}
},
{
"timestamp_unix_nano": 6000000000,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "x",
"attributes": {
"cairnobs.aggregated": "true",
"cairnobs.count": "1",
"cairnobs.window_start_unix_nano": "6000000000",
"cairnobs.window_last_unix_nano": "6000000000"
}
}
]
}
processing/conformance/cases/aggregate_count_collapses_a_window_into_one_tagged_record.json
+56
View File
@@ -0,0 +1,56 @@
{
"name": "aggregate_count_collapses_a_window_into_one_tagged_record",
"description": "The emitted record is the window's first record unchanged, plus the aggregate attributes. Keeping the first record means a reader sees a real example of what was collapsed rather than an invented summary.",
"rules": [
{
"match": [],
"actions": [
{
"action": "aggregate_count",
"window_ms": 5000
}
]
}
],
"inputs": [
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "repeated",
"attributes": {}
},
{
"timestamp_unix_nano": 1000000000,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "repeated",
"attributes": {}
},
{
"timestamp_unix_nano": 2000000000,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "repeated",
"attributes": {}
}
],
"expect": [
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "repeated",
"attributes": {
"cairnobs.aggregated": "true",
"cairnobs.count": "3",
"cairnobs.window_start_unix_nano": "0",
"cairnobs.window_last_unix_nano": "2000000000"
}
}
]
}
processing/conformance/cases/aggregate_count_honours_key_fields.json
+73
View File
@@ -0,0 +1,73 @@
{
"name": "aggregate_count_honours_key_fields",
"description": "With key_fields, each key gets its own window and its own aggregate. Emission order at end of stream follows the order the windows were opened.",
"rules": [
{
"match": [],
"actions": [
{
"action": "aggregate_count",
"window_ms": 5000,
"key_fields": [
"host",
"message"
]
}
]
}
],
"inputs": [
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "z",
"attributes": {}
},
{
"timestamp_unix_nano": 100000000,
"host": "h2",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "z",
"attributes": {}
},
{
"timestamp_unix_nano": 200000000,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "z",
"attributes": {}
}
],
"expect": [
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "z",
"attributes": {
"cairnobs.aggregated": "true",
"cairnobs.count": "2",
"cairnobs.window_start_unix_nano": "0",
"cairnobs.window_last_unix_nano": "200000000"
}
},
{
"timestamp_unix_nano": 100000000,
"host": "h2",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "z",
"attributes": {
"cairnobs.aggregated": "true",
"cairnobs.count": "1",
"cairnobs.window_start_unix_nano": "100000000",
"cairnobs.window_last_unix_nano": "100000000"
}
}
]
}
processing/conformance/cases/aggregate_count_leaves_non_matching_records_alone.json
+70
View File
@@ -0,0 +1,70 @@
{
"name": "aggregate_count_leaves_non_matching_records_alone",
"description": "Records the rule did not match pass through untouched and in order, interleaved with the aggregate the matched ones produced.",
"rules": [
{
"match": [
{
"field": "service",
"op": "eq",
"value": "noisy"
}
],
"actions": [
{
"action": "aggregate_count",
"window_ms": 5000
}
]
}
],
"inputs": [
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "noisy",
"severity": "SEVERITY_INFO",
"message": "n",
"attributes": {}
},
{
"timestamp_unix_nano": 100000000,
"host": "h1",
"service": "quiet",
"severity": "SEVERITY_INFO",
"message": "keep me",
"attributes": {}
},
{
"timestamp_unix_nano": 200000000,
"host": "h1",
"service": "noisy",
"severity": "SEVERITY_INFO",
"message": "n",
"attributes": {}
}
],
"expect": [
{
"timestamp_unix_nano": 100000000,
"host": "h1",
"service": "quiet",
"severity": "SEVERITY_INFO",
"message": "keep me",
"attributes": {}
},
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "n",
"attributes": {
"cairnobs.aggregated": "true",
"cairnobs.count": "2",
"cairnobs.window_start_unix_nano": "0",
"cairnobs.window_last_unix_nano": "200000000"
}
}
]
}
processing/conformance/cases/aggregate_count_tags_even_a_single_occurrence.json
+40
View File
@@ -0,0 +1,40 @@
{
"name": "aggregate_count_tags_even_a_single_occurrence",
"description": "A window that saw one record is still tagged with count 1. Emitting a bare record here would make cairnobs.count present only sometimes, which silently breaks summing it on quiet windows.",
"rules": [
{
"match": [],
"actions": [
{
"action": "aggregate_count",
"window_ms": 5000
}
]
}
],
"inputs": [
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "lonely",
"attributes": {}
}
],
"expect": [
{
"timestamp_unix_nano": 0,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "lonely",
"attributes": {
"cairnobs.aggregated": "true",
"cairnobs.count": "1",
"cairnobs.window_start_unix_nano": "0",
"cairnobs.window_last_unix_nano": "0"
}
}
]
}
processing/conformance/cases/aggregate_count_window_bounds_are_observed_not_nominal.json
+48
View File
@@ -0,0 +1,48 @@
{
"name": "aggregate_count_window_bounds_are_observed_not_nominal",
"description": "window_last is the last record that actually contributed, never window_start + window_ms. A window flushed early must not claim an end that never happened.",
"rules": [
{
"match": [],
"actions": [
{
"action": "aggregate_count",
"window_ms": 60000
}
]
}
],
"inputs": [
{
"timestamp_unix_nano": 100000000,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "y",
"attributes": {}
},
{
"timestamp_unix_nano": 250000000,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "y",
"attributes": {}
}
],
"expect": [
{
"timestamp_unix_nano": 100000000,
"host": "h1",
"service": "s1",
"severity": "SEVERITY_INFO",
"message": "y",
"attributes": {
"cairnobs.aggregated": "true",
"cairnobs.count": "2",
"cairnobs.window_start_unix_nano": "100000000",
"cairnobs.window_last_unix_nano": "250000000"
}
}
]
}
processing/conformance/validate.py
+8 -5
View File
@@ -45,12 +45,15 @@ ACTIONS = {
"parse_regex": ({"field", "pattern"}, set()),
"sample": ({"keep_one_in"}, set()),
"suppress_duplicates": ({"window_ms"}, {"key_fields"}),
"aggregate_count": ({"window_ms"}, {"key_fields"}),
}
# Deliberately absent from ACTIONS. The design does not say what it emits
# or what a query that is not expecting a synthetic record sees, so a
# case using it would invent that answer and freeze it by accident.
UNSPECIFIED_ACTIONS = {"aggregate_count"}
# Actions the design names but does not yet specify the output of. A case
# using one would invent that answer and freeze it by accident, so the
# validator refuses them until the design says what they emit.
# aggregate_count was here until 2026-09-05; see Decision 5 in
# /docs/phase-8-processing-design.md.
UNSPECIFIED_ACTIONS = set()
def valid_field(name):
@@ -120,7 +123,7 @@ def check_action(a, where, err):
n = a.get("keep_one_in")
if not isinstance(n, int) or n < 1:
err(f"{where}: keep_one_in must be an integer >= 1")
if name == "suppress_duplicates":
if name in ("suppress_duplicates", "aggregate_count"):
w = a.get("window_ms")
if not isinstance(w, int) or w < 1:
err(f"{where}: window_ms must be an integer >= 1")