microshift-release-pre-check

name	microshift-release:pre-check
description	Check OCP release schedule, verify availability, evaluate z-stream need, or check nightly build gaps

microshift-release:pre-check

Synopsis

/microshift-release:pre-check [release_type] [version|time-range...] [--verbose]

Description

MicroShift ships as a layered product on top of OCP. Every time OCP releases a new version (z-stream, EC, RC, or nightly), the MicroShift team must evaluate whether to participate — checking for CVEs, verifying RPM builds exist in Brew, and deciding whether to ask ART to create artifacts.

This command automates that evaluation (Phase 0 of the release process). It checks lifecycle status, OCP payload availability, advisory CVEs, nightly build gaps, and EC/RC readiness — then outputs a clear action per version: OK, SKIP, ASK ART, NEEDS REVIEW, or ALREADY RELEASED.

When a time range is provided (e.g., "this week"), it queries ART Jira for OCP release tickets due in that period and evaluates each one.

Prerequisites

Requirement	Needed for	Mandatory?
VPN	Brew RPM checks (nightly, EC/RC), advisory report	Yes for nightly/ecrc — xyz degrades gracefully (skips advisory, 90-day rule)
Atlassian MCP	OCPBUGS enrichment, ART ticket queries, time range lookups	Yes — required to analyze OCPBUGS resolution and release action
GITLAB_API_TOKEN	Advisory report for 4.20+ (shipment MR data)	No — advisory skipped for 4.20+ without it

Arguments

• release_type (optional): One or more of Z, X, Y, RC, EC, nightly (case-insensitive). If omitted, defaults to Z.
• version (optional): Specific version (e.g., 4.19.27) or minor stream (e.g., 4.21)
• time-range (optional): Natural language time range instead of explicit versions. Detected by keywords like:
  • today, tomorrow
  • this week, next week
  • next 3 days, next 7 days
  • this month
• --verbose (optional): Show extra detail (tables for xyz, NVR/nightly names for nightly, next versions for EC/RC).

Scripts Directory

All scripts are run relative to the repository root:

SCRIPTS_DIR=plugins/microshift-release/scripts

Implementation

Step 1: Parse Arguments

1. Identify release_type(s) — tokens matching Z, X, Y, RC, EC, nightly (case-insensitive)
2. Identify version(s) — tokens matching X.Y or X.Y.Z pattern
3. Identify time range — remaining tokens that are not release types, versions, or flags (e.g., "this week", "next 3 days", "tomorrow")
4. Identify --verbose flag
5. Default: If no release_type found, default to Z and treat version/time-range tokens accordingly

Step 2: Resolve Versions via ART Jira (when time range is detected)

If a time range is present instead of explicit versions, query ART Jira for release tickets due in that window:

1. Convert the time range to concrete dates (date_from, date_to) based on today's date:

   • today → today only
   • tomorrow → tomorrow only
   • this week → Monday through Sunday of the current week
   • next week → next Monday through next Sunday
   • next N days → today through N days from now
   • this month → today through end of current month
   • For any other natural language range, compute the appropriate date window

2. Query ART Jira using mcp__atlassian__searchJiraIssuesUsingJql:

   JQL: project = ART AND issuetype = Story AND summary ~ "Release 4." AND duedate >= "{date_from}" AND duedate <= "{date_to}" ORDER BY duedate ASC
   

   Use cloudId: "redhat.atlassian.net" for the Atlassian MCP tool.

3. Extract versions from ticket summaries. ART release tickets use the format "Release X.Y.Z [YYYY-Mon-DD]" (e.g., "Release 4.21.18 [2026-Jun-02]"). Extract the X.Y.Z version from each matching ticket.

4. Filter to 4.14+ only (MicroShift GA'd at 4.14 — older versions have no MicroShift images).

5. Pass the resolved versions as explicit arguments to the script in Step 4.

If no ART tickets are found in the date range, report "No OCP releases scheduled in {range}."

Step 3: Query ART Tickets via MCP

Before running the script, query ART Jira for in-progress release tickets so the script can show ART ticket status in the Release Schedule table.

1. Call mcp__atlassian__searchJiraIssuesUsingJql with:

   • cloudId: "redhat.atlassian.net"
   • jql: project = ART AND summary ~ "Release" AND status = "In Progress" ORDER BY duedate ASC
   • fields: ["summary", "status", "duedate"]
   • maxResults: 50

2. From the results, build a JSON array:

   [{"key": "ART-XXXXX", "summary": "Release 4.21.18 [2026-Jun-03]", "status": "In Progress", "due_date": "2026-06-03"}]
   

3. Write the JSON to a temp file and set ART_TICKETS_JSON env var:

   echo '<json>' > /tmp/art_tickets.json
   

If mcp__atlassian__searchJiraIssuesUsingJql is not available, skip this step — the script degrades gracefully (shows None for ART tickets).

Step 4: Run the Script

Map each release type to the corresponding precheck.sh subcommand and run via Bash:

Release Type	Command
Z, X, Y (default)	ART_TICKETS_JSON=/tmp/art_tickets.json bash ${SCRIPTS_DIR}/precheck.sh xyz [versions...]
nightly	bash ${SCRIPTS_DIR}/precheck.sh nightly [version]
EC	ART_TICKETS_JSON=/tmp/art_tickets.json bash ${SCRIPTS_DIR}/precheck.sh ecrc EC [version]
RC	ART_TICKETS_JSON=/tmp/art_tickets.json bash ${SCRIPTS_DIR}/precheck.sh ecrc RC [version]

IMPORTANT: Only append --verbose to the command if the user explicitly passed --verbose in their arguments. Do NOT add it by default.

Stderr contains progress messages — only display it if the script exits non-zero.

Multiple types (e.g., nightly EC RC): Run each command as a separate Bash call in parallel.

Step 5: Display Output

Display the script output verbatim — do not reformat, add tables, or change the layout. The scripts produce deterministic pre-formatted text. Do NOT add any commentary, explanation, or summary after the output.

OCPBUGS follow-up: If any version shows OCPBUGS in the output (e.g., 1 OCPBUGS), automatically re-run the command with --verbose to list the specific bugs. Only do this once — do not re-run if --verbose was already passed.

Step 6: Enrich OCPBUGS via MCP

After displaying the output (including any --verbose re-run), if any OCPBUGS appeared in the results:

1. Collect OCPBUGS keys: Extract all unique OCPBUGS-XXXXX keys from the output (they appear in the Resolved OCPBUGS table or the one-line summaries).

2. Fetch each bug via MCP: For each unique key, call mcp__atlassian__getJiraIssue with:

   • cloudId: "redhat.atlassian.net"
   • issueIdOrKey: the OCPBUGS key (e.g., "OCPBUGS-12345")
   • fields: ["summary", "status", "labels", "issuetype", "priority"]
   • responseContentFormat: "markdown" Make all getJiraIssue calls in parallel (multiple tool calls in one message).

3. Build enriched JSON: For each successfully fetched bug, build a JSON object:

   {
     "key": "OCPBUGS-12345",
     "version": "4.21",
     "summary": "<from MCP response: fields.summary>",
     "status": "<from MCP response: fields.status.name>",
     "labels": ["<from MCP response: fields.labels array>"],
     "issuetype": "<from MCP response: fields.issuetype.name>",
     "priority": "<from MCP response: fields.priority.name>"
   }
   

   The version field is the minor version (e.g., "4.21") from the evaluation that referenced the bug. If a bug appears in multiple versions, include one entry per version.

4. Render enrichment table: Pipe the JSON array through the enrichment script:

   echo '<json_array>' | bash ${SCRIPTS_DIR}/precheck.sh enrich
   

5. Display the enrichment output after the main precheck output. This shows real summaries, statuses, release actions (release-required/release-not-required/needs-review), and updated recommendations.

If mcp__atlassian__getJiraIssue is not available, skip enrichment and note that the Atlassian MCP is required for OCPBUGS analysis.

Step 7: Handle Errors

If the script exits non-zero, display stderr and suggest:

• VPN not connected → connect to VPN (Brew requires it)
• Missing env vars → set GITLAB_API_TOKEN (for 4.20+ advisory reports)

Examples

/microshift-release:pre-check this week                   # OCP versions releasing this week
/microshift-release:pre-check next week                   # OCP versions releasing next week
/microshift-release:pre-check today                       # OCP versions releasing today
/microshift-release:pre-check next 3 days                 # OCP versions in next 3 days
/microshift-release:pre-check 4.21.10                     # specific version
/microshift-release:pre-check 4.20 4.21 4.22              # xyz eval for multiple streams
/microshift-release:pre-check 4.19.27 --verbose           # specific version, detailed report
/microshift-release:pre-check nightly                     # nightly gaps for all active branches
/microshift-release:pre-check EC                          # latest EC status
/microshift-release:pre-check RC                          # latest RC status
/microshift-release:pre-check nightly EC RC               # combined report

Notes

• Read-only — does NOT create tickets or modify external state
• Scripts support --json for raw JSON output when called directly (e.g., bash ${SCRIPTS_DIR}/precheck.sh xyz 4.21.10 --json)
• --verbose works for all types: detailed tables for xyz, NVR/nightly names for nightly, next versions for EC/RC
• OCPBUGS enrichment uses Atlassian MCP (OAuth) — no PAT env vars needed; the script discovers bug keys from git commits, and the skill enriches them via getJiraIssue
• VPN required for Brew and errata access