Give each tracker agent its own config

The tracker-mapping, common-pattern enrichment, and third-party
disambiguation agents were all built from one shared
TrackerAgentsConfig fed by a single tracker-mapping config slot. That
forced a single AgentTimeout to be reused and patched per worker, and
two unrelated max-turns fields to share one struct.

Split the in-code config into TrackerMappingAgentConfig,
TrackerEnrichmentAgentConfig, and DisambiguationAgentConfig, each with
its own timeout and max-turns, and add dedicated tracker-enrichment and
third-party-disambiguation provider slots (the latter resolving next to
third-party-vetter). Enrichment and disambiguation fall back to the
tracker-mapping slot when their own provider is unset, preserving
single-config deployments.

Drop the shared pkg/agentsbuild package and duplicate its small wiring
into probod and proboctl so the two executables stay decoupled. Wire
the new env vars, builder test coverage, and Helm values.

Signed-off-by: Émile Ré <emile@probo.com>
This commit is contained in:
Émile Ré
2026-06-08 11:24:37 +02:00
parent e6866f88a0
commit 92a73e9302
20 changed files with 552 additions and 289 deletions

View File

@@ -281,6 +281,23 @@ spec:
- name: THIRD_PARTY_VETTING_STALE_AFTER
value: {{ .Values.probo.thirdPartyVettingWorker.staleAfter | quote }}
{{- end }}
# Third-party Disambiguation Agent
{{- if .Values.probo.thirdPartyDisambiguation.provider }}
- name: AGENT_THIRD_PARTY_DISAMBIGUATION_PROVIDER
value: {{ .Values.probo.thirdPartyDisambiguation.provider | quote }}
{{- end }}
{{- if .Values.probo.thirdPartyDisambiguation.modelName }}
- name: AGENT_THIRD_PARTY_DISAMBIGUATION_MODEL_NAME
value: {{ .Values.probo.thirdPartyDisambiguation.modelName | quote }}
{{- end }}
{{- if .Values.probo.thirdPartyDisambiguation.temperature }}
- name: AGENT_THIRD_PARTY_DISAMBIGUATION_TEMPERATURE
value: {{ .Values.probo.thirdPartyDisambiguation.temperature | quote }}
{{- end }}
{{- if .Values.probo.thirdPartyDisambiguation.maxTokens }}
- name: AGENT_THIRD_PARTY_DISAMBIGUATION_MAX_TOKENS
value: {{ .Values.probo.thirdPartyDisambiguation.maxTokens | quote }}
{{- end }}
# Tracker Mapping Agent
{{- if .Values.probo.trackerMapping.provider }}
- name: AGENT_TRACKER_MAPPING_PROVIDER
@@ -298,6 +315,23 @@ spec:
- name: AGENT_TRACKER_MAPPING_MAX_TOKENS
value: {{ .Values.probo.trackerMapping.maxTokens | quote }}
{{- end }}
# Tracker Enrichment Agent
{{- if .Values.probo.trackerEnrichment.provider }}
- name: AGENT_TRACKER_ENRICHMENT_PROVIDER
value: {{ .Values.probo.trackerEnrichment.provider | quote }}
{{- end }}
{{- if .Values.probo.trackerEnrichment.modelName }}
- name: AGENT_TRACKER_ENRICHMENT_MODEL_NAME
value: {{ .Values.probo.trackerEnrichment.modelName | quote }}
{{- end }}
{{- if .Values.probo.trackerEnrichment.temperature }}
- name: AGENT_TRACKER_ENRICHMENT_TEMPERATURE
value: {{ .Values.probo.trackerEnrichment.temperature | quote }}
{{- end }}
{{- if .Values.probo.trackerEnrichment.maxTokens }}
- name: AGENT_TRACKER_ENRICHMENT_MAX_TOKENS
value: {{ .Values.probo.trackerEnrichment.maxTokens | quote }}
{{- end }}
# Tracker Mapping Worker
{{- if .Values.probo.trackerMappingWorker.interval }}
- name: TRACKER_MAPPING_INTERVAL
@@ -319,6 +353,10 @@ spec:
- name: TRACKER_MAPPING_AGENT_MAX_TURNS
value: {{ .Values.probo.trackerMappingWorker.agentMaxTurns | quote }}
{{- end }}
{{- if .Values.probo.trackerMappingWorker.disambiguationAgentTimeout }}
- name: TRACKER_MAPPING_DISAMBIGUATION_AGENT_TIMEOUT
value: {{ .Values.probo.trackerMappingWorker.disambiguationAgentTimeout | quote }}
{{- end }}
# Common Pattern Enrichment Worker
{{- if .Values.probo.commonPatternEnrichmentWorker.interval }}
- name: COMMON_PATTERN_ENRICHMENT_INTERVAL

View File

@@ -185,6 +185,15 @@ probo:
# maxConcurrency: 1
# staleAfter: 1500
# Third-party disambiguation agent (optional; runs inside the tracker
# mapping worker to pick the best matching org third party). Falls back
# to trackerMapping when its provider is unset.
# thirdPartyDisambiguation:
# provider: "openai"
# modelName: "gpt-4o"
# temperature: "0.4"
# maxTokens: "4096"
# Tracker mapping agent (optional, auto-links tracker patterns to vendors)
# trackerMapping:
# provider: "openai"
@@ -192,8 +201,16 @@ probo:
# temperature: "0.1"
# maxTokens: "4096"
# Tracker enrichment agent (optional; researches catalog descriptions).
# Falls back to trackerMapping when its provider is unset.
# trackerEnrichment:
# provider: "openai"
# modelName: "gpt-4o"
# temperature: "0.2"
# maxTokens: "4096"
# Tracker mapping worker tuning (optional; seconds for
# interval/staleAfter/agentTimeout).
# interval/staleAfter/agentTimeout/disambiguationAgentTimeout).
# Keep concurrency modest to stay under OpenAI/Firecrawl limits and the DB pool.
# trackerMappingWorker:
# interval: 10
@@ -201,6 +218,7 @@ probo:
# staleAfter: 600
# agentTimeout: 45
# agentMaxTurns: 10
# disambiguationAgentTimeout: 45
# Common-pattern enrichment worker tuning (optional; seconds for
# interval/staleAfter/agentTimeout).

View File

@@ -286,6 +286,15 @@ probo:
maxConcurrency: 1
staleAfter: 1500
# Third-party disambiguation agent (optional; runs inside the tracker
# mapping worker to pick the best matching org third party). Falls back
# to trackerMapping when its provider is unset.
thirdPartyDisambiguation:
provider: ""
modelName: ""
temperature: ""
maxTokens: ""
# Tracker mapping agent (optional, requires openai.apiKey or anthropic key)
trackerMapping:
provider: ""
@@ -293,16 +302,25 @@ probo:
temperature: ""
maxTokens: ""
# Tracker enrichment agent (optional; researches catalog descriptions).
# Falls back to trackerMapping when its provider is unset.
trackerEnrichment:
provider: ""
modelName: ""
temperature: ""
maxTokens: ""
# Tracker mapping background worker tuning (optional). interval,
# staleAfter, and agentTimeout are in seconds. Keep concurrency modest
# to stay under OpenAI/Firecrawl rate limits and the database
# connection pool.
# staleAfter, agentTimeout, and disambiguationAgentTimeout are in
# seconds. Keep concurrency modest to stay under OpenAI/Firecrawl rate
# limits and the database connection pool.
trackerMappingWorker:
interval: 10
maxConcurrency: 3
staleAfter: 600
agentTimeout: 45
agentMaxTurns: 10
disambiguationAgentTimeout: 45
# Common-pattern enrichment background worker tuning (optional).
# interval, staleAfter, and agentTimeout are in seconds.