Forge
markdowne8ad0934
1# Linux Environments Playbook v1
2
3## Purpose
4Project-independent playbook for reliable engineering and troubleshooting on Linux hosts.
5
6## Scope
7- Linux process/service runtime behavior
8- Permissions and ownership model
9- Resource limits and isolation boundaries
10- Operational diagnostics baseline
11
12## Evidence-Based Working Format
13- Fact: capture concrete host symptom (service crash loop, permission issue, limit exhaustion).
14- Hypothesis: define smallest plausible Linux-level cause.
15- Check: run minimal reproducible verification command set.
16- Decision criterion: predefine close/escalate threshold.
17- Confidence mark: explicit certainty and residual risk.
18
19## Core Linux Contracts
20- Service model and lifecycle semantics must be explicit.
21- Ownership/mode/capability assumptions must be validated.
22- Resource limits and cgroup boundaries must be visible.
23- Locale/timezone/encoding assumptions must be explicit.
24
25## Runtime and Service Contracts
26- systemd unit contracts (dependencies, restart, environment) must be deterministic.
27- Startup arguments and runtime env must be explicit.
28- Signal handling and graceful shutdown behavior must be validated.
29- Worker/daemon ownership of subprocesses must be clear.
30
31## Permissions and Filesystem Contracts
32- Validate uid/gid ownership and mode bits on critical paths.
33- Keep mount and symlink boundaries explicit.
34- Avoid implicit root assumptions.
35- Confirm write boundaries before mutating operations.
36
37## Resource and Isolation Contracts
38- Check `ulimit` and file descriptor ceilings for IO-heavy services.
39- Validate memory/cpu/pid constraints in cgroups/containers.
40- Distinguish host saturation from app-level faults.
41- Keep capacity symptoms separated from logic bugs.
42
43## Diagnostics Baseline
44- journal logs + structured application logs.
45- Process tree, open-file/socket state, and limit visibility.
46- Resource pressure timeline (cpu/memory/io) before deep app debugging.
47- One concise root-cause and prevention delta per incident.
48
49## Metrics
50- Mean time to identify Linux host root cause.
51- Frequency of limit/ownership-related incidents.
52- Recovery time for service restarts.
53- Reopen rate for Linux environment incidents.
54
55## Revisit Triggers
56- Repeated “works only on one distro/host profile” failures.
57- Frequent permission/ownership regressions.
58- Recurring incidents from invisible resource constraints.
59
60
View only · write via MCP/CIDE