oci-shepherd/config.example.yaml at master · slammingprogramming/oci-shepherd · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
# oci-shepherd configuration
#
# Copy this file to config.yaml (or /etc/oci-shepherd/config.yaml
# for the systemd install) and fill in your own tenancy's values. Every
# OCID below is a fake placeholder - replace them all.
#
# NOTE: when uncommenting an optional line below, delete only the leading
# '#' character. If a stray space is left in front of the key (e.g.
# " image_id: ..." instead of "image_id: ..."), YAML will fail to parse
# with an indentation error.
#
# This tool does NOT read or manage OCI credentials. It calls the `oci` CLI,
# which must already be authenticated on this machine (`~/.oci/config`, or
# OCI_CLI_* environment variables) exactly as if you were running `oci` by
# hand. Test that first:
#     oci iam availability-domain list --compartment-id <your compartment OCID>

# --- Identity / region -------------------------------------------------

# The OCI region to operate in, e.g. us-ashburn-1, uk-london-1, eu-frankfurt-1.
region: us-ashburn-1

# Optional: an OCI CLI profile name from ~/.oci/config to use instead of
# DEFAULT. Leave unset (or remove the line) to use the CLI's default profile.
# profile: DEFAULT

# The compartment OCID to launch the instance into. To launch into the
# tenancy's root compartment, use your tenancy OCID here instead
# (starts with "ocid1.tenancy." rather than "ocid1.compartment.") - OCI
# treats the tenancy OCID as the root compartment's identifier.
compartment_id: ocid1.compartment.oc1..aaaaaaaaexampleexampleexampleexampleexampleexampleexampleexa

# --- Shape ---------------------------------------------------------------

# Always Free eligible shapes:
#   VM.Standard.E2.1.Micro   - AMD, fixed 1 OCPU / 1 GB RAM, no shape_config needed
#   VM.Standard.A1.Flex      - Ampere ARM, flexible OCPU/RAM, requires shape_config
#                              below (Always Free limit: up to 4 OCPUs / 24 GB RAM
#                              total across all A1 instances in a tenancy)
shape: VM.Standard.A1.Flex

# Required only when shape ends in ".Flex". Stay within your tenancy's
# Always Free A1 limits (commonly 4 OCPUs / 24 GB total).
shape_config:
  ocpus: 4
  memory_in_gbs: 24

# --- Availability domains -------------------------------------------------

# Which ADs in the target region to cycle through between launch attempts.
# Capacity errors are AD-specific, so trying every AD in the region gives
# the best chance of success. Use ["all"] to auto-discover every AD in the
# region/compartment, or list specific ADs explicitly, e.g.:
#   availability_domains:
#     - "abCD:US-ASHBURN-AD-1"
#     - "abCD:US-ASHBURN-AD-2"
#     - "abCD:US-ASHBURN-AD-3"
availability_domains:
  - all

# --- Networking ------------------------------------------------------------

# Subnet OCID to attach the instance's primary VNIC to. This tool does not
# create VCNs/subnets - create one first (Always Free includes 2 VCNs).
subnet_id: ocid1.subnet.oc1..aaaaaaaaexampleexampleexampleexampleexampleexampleexampleexampl

# Whether to assign a public (ephemeral) IP to the instance. true/false.
assign_public_ip: true

# --- Image -----------------------------------------------------------------

# Option A: pin an exact image OCID (recommended for reproducibility).
# image_id: ocid1.image.oc1..aaaaaaaaexampleexampleexampleexampleexampleexampleexampleexamp

# Option B: leave image_id unset/commented and resolve the newest matching
# image by OS name + version instead. Both fields below are required if
# image_id is not set. Run `oci compute image list --compartment-id <ocid>
# --shape <shape>` to see what's available in your region.
image_operating_system: "Canonical Ubuntu"
image_operating_system_version: "22.04"

# --- SSH access --------------------------------------------------------

# Path to a public key file (recommended), OR set ssh_public_key inline
# instead. Exactly one of the two is required.
ssh_public_key_path: ~/.ssh/id_ed25519.pub
# ssh_public_key: "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAI... you@example.com"

# --- Instance ----------------------------------------------------------

# Display name given to the launched instance.
display_name: always-free-instance

# Boot volume size in GB. Always Free total boot volume allowance is 200 GB
# across all boot + block volumes combined.
boot_volume_size_in_gbs: 50

# --- Timing / retry behavior ---------------------------------------------

# How long to wait, in seconds, after a full pass through every availability
# domain before starting another pass (obtain mode). Community reports for
# A1 capacity commonly range from minutes to weeks depending on region/AD,
# so this can safely be a few minutes without meaningfully slowing things
# down - see README.md's troubleshooting section.
obtain_retry_interval_seconds: 60

# How long to wait between trying each individual AD within a single pass.
obtain_inter_ad_delay_seconds: 5

# How long, in seconds, to wait for the instance to reach RUNNING state
# after a launch call succeeds before giving up on that attempt.
wait_for_running_seconds: 600

# How often, in seconds, monitor mode polls the instance's state once obtained.
monitor_check_interval_seconds: 300

# --- Rate limit (HTTP 429) backoff ----------------------------------------

# Base and max delay, in seconds, for exponential backoff (with jitter) when
# the OCI API returns HTTP 429 Too Many Requests. Delay doubles on each
# consecutive 429 for the same availability domain, up to the max, then a
# random 0-30% jitter is added. This is separate from obtain_retry_interval_seconds,
# which governs normal "no capacity" retries.
rate_limit_backoff_base_seconds: 30
rate_limit_backoff_max_seconds: 900

# How many consecutive 429s on the same availability domain to tolerate
# (backing off between each) before giving up on that AD for this pass and
# moving on to the next one. Optional, defaults to 8.
rate_limit_max_consecutive_retries: 8

# --- Logging / state -----------------------------------------------------

# Structured JSON Lines log of every attempt (timestamp, AD, shape, outcome,
# error detail). The parent directory is created automatically if missing.
log_file: /var/log/oci-shepherd/activity.jsonl

# Where the tool records the obtained instance's OCID so monitor mode (a
# separate process) knows what to watch. The parent directory is created
# automatically if missing.
state_file: /var/lib/oci-shepherd/state.json

# --- systemd handoff -------------------------------------------------------

# Unit names this tool runs `systemctl start <unit>` on to hand off between
# modes. Only relevant if you used different unit names than the ones
# shipped in systemd/. Change these if you renamed the unit files.
obtain_systemd_unit: oci-shepherd-obtain.service
monitor_systemd_unit: oci-shepherd-monitor.service