unknown.yaml

# Identifiers seen in the Translator platform that no component file claims.
#
# A holding pen, not a component. Anything observed in the wild — an
# OpenTelemetry service name, a hostname, a registry entry — that we cannot
# attribute to a components/*.yaml goes here rather than being dropped, so the
# next person does not have to rediscover it.
#
# Entries leave this file in one of two ways:
#   - promoted, when we learn which component it belongs to: move the
#     identifier into that component's file (or write a new component file for
#     it) and delete the entry here;
#   - retired, when someone confirms it is no longer in use: set
#     `status: not-in-use` and leave it, so it is not investigated twice.
#
# `status` values:
#   unattributed  we do not know what this is
#   operation     a processing step reported as its own service, not a component
#   not-recorded  a known component that has no file yet, named by `component`
#   not-in-use    confirmed retired
#
# Validated by schema/unknown.schema.json; see tests/test_components.py.
# New kinds of unattributed identifier get a new top-level key (`urls:` is
# already in the schema, and is empty because nothing has needed it yet).
# yaml-language-server: $schema=schema/unknown.schema.json

otel_services:

  # --- Genuinely unattributed -------------------------------------------------

  - name: shepherd-server
    status: unattributed
    first_seen: "2026-08-31"
    environments: [ci, test]
    component: null
    evidence: >-
      Parent span of every per-ARA Shepherd worker (aragorn, arax, bte) and of
      every operation below, and a child of ARS. So it is the shared Shepherd
      server rather than any one of shepherd-arax / shepherd-aragorn /
      shepherd-bte, and it probably wants a component file of its own — the
      three shepherd-* files all point at BioPack-team/shepherd and the
      `shepherd` Helm chart, and none of them is this.
    note: >-
      Confirmed complex and deliberately left here on 2026-08-31 rather than
      forced into one of the shepherd-* files. Re-investigating needs a
      decision about whether the shared server is its own component, not more
      trace data.

  - name: biothings-plover
    status: unattributed
    first_seen: "2026-08-31"
    environments: [ci, test]
    component: null
    evidence: >-
      A Plover instance that is not ploverdb: rtx-kg2-plover is the one
      components/ploverdb.yaml claims, and multiomics-plover is a separate KP.
      Not seen in any trace collected so far, so there is no call-graph
      evidence for who talks to it.

  # --- Processing steps, not components ---------------------------------------
  #
  # Each of these is a child span of shepherd-server and reports under its own
  # service name, so the collector lists them alongside real services. They are
  # Shepherd operations. Listed here so nobody attributes them to a component.

  - name: filter_analyses_top_n
    status: operation
    first_seen: "2026-08-31"
    environments: [ci, test]
    component: shepherd-server
  - name: filter_kgraph_orphans
    status: operation
    first_seen: "2026-08-31"
    environments: [ci, test]
    component: shepherd-server
  - name: filter_results_top_n
    status: operation
    first_seen: "2026-08-31"
    environments: [ci, test]
    component: shepherd-server
  - name: finish_query
    status: operation
    first_seen: "2026-08-31"
    environments: [ci, test]
    component: shepherd-server
    note: Reports ARS as a child span, so it is where results go back.
  - name: merge_message
    status: operation
    first_seen: "2026-08-31"
    environments: [ci, test]
    component: shepherd-server
  - name: score_paths
    status: operation
    first_seen: "2026-08-31"
    environments: [ci, test]
    component: shepherd-server
  - name: sort_results_score
    status: operation
    first_seen: "2026-08-31"
    environments: [ci, test]
    component: shepherd-server

  # --- Known components with no file yet --------------------------------------
  #
  # These are rows in the sheet whose refactor status keeps them out of the 26
  # files recorded so far. They are listed so the OTel service list can be
  # reconciled completely, and they get files when we convert the rest.

  - name: aragorn-aragorn-server-ARAGORN
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [prod]
    component: aragorn
    evidence: >-
      The standalone ARAGORN ARA. Distinct from the `aragorn` service in
      ci/test, which is a child of shepherd-server and is claimed by
      components/shepherd-aragorn.yaml.
  - name: automat-cam-kp
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [test, prod]
    component: automat
  - name: automat-hgnc
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [test, prod]
    component: automat
  - name: automat-icees-kg
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [test, prod]
    component: automat
  - name: automat-monarch-kg
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [test, prod]
    component: automat
  - name: automat-robokopkg
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [test, prod]
    component: automat
  - name: biothings-explorer
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [test]
    component: biothings-explorer
    evidence: >-
      The standalone BTE. Distinct from the `bte` service in ci/test, which is
      a child of shepherd-server and is claimed by components/shepherd-bte.yaml.
  - name: COHD
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [prod]
    component: cohd-api
  - name: genetics-data-provider
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [test, prod]
    component: genetics-kp
  - name: multiomics-plover
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [prod]
    component: multiomics
  - name: strider
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [prod]
    component: strider
  - name: STRIDER-DEV
    status: not-recorded
    first_seen: "2026-08-31"
    environments: [test]
    component: strider
    note: A second service name for the same component, in the test collector.

  # --- No sheet row at all ----------------------------------------------------

  - name: molepro
    status: unattributed
    first_seen: "2026-08-31"
    environments: [prod]
    component: null
    evidence: >-
      The Molecular Data Provider KP. The sheet has moleproapi-pipeline and
      moleprodb-pipeline but no row for the service itself, so there is no
      component id to attribute this to yet.
    note: >-
      Confirmed complex and deliberately left here on 2026-08-31. Resolving it
      needs a decision about how MolePro's pipelines and its service relate,
      not more trace data.

# Nothing has reached `status: not-in-use` yet: every name above was reporting
# to a live collector on 2026-08-31.
#
# Promoted out of this file so far:
#   NodeNorm -> components/nodenorm-es.yaml (2026-08-31). Its spans carry
#   http.host=nodenorm-es.{ci,test}.transltr.io and a Tornado handler from
#   the nodenorm package, backed by elasticsearch.es-core-components — so
#   it is the ES rewrite, not the nodenorm it replaced.