misc/scratch: derive AMIs by architecture and add instance-type override

ci/builder/Dockerfile

@@ -172,7 +172,7 @@ COPY nodesource.asc .
 
 RUN gpg --dearmor < nodesource.asc > /etc/apt/keyrings/nodesource.gpg \
     && rm nodesource.asc \
-    && echo "deb [signed-by=/etc/apt/keyrings/nodesource.gpg] https://deb.nodesource.com/node_22.x nodistro main" > /etc/apt/sources.list.d/nodesource.list \
+    && echo "deb [signed-by=/etc/apt/keyrings/nodesource.gpg] https://deb.nodesource.com/node_24.x nodistro main" > /etc/apt/sources.list.d/nodesource.list \
     && apt-get update \
     && apt-get install -y --no-install-recommends nodejs \
     && apt-get clean \
@@ -496,7 +496,7 @@ COPY nodesource.asc .
 
 RUN gpg --dearmor < nodesource.asc > /etc/apt/keyrings/nodesource.gpg \
     && rm nodesource.asc \
-    && echo "deb [signed-by=/etc/apt/keyrings/nodesource.gpg] https://deb.nodesource.com/node_22.x nodistro main" > /etc/apt/sources.list.d/nodesource.list \
+    && echo "deb [signed-by=/etc/apt/keyrings/nodesource.gpg] https://deb.nodesource.com/node_24.x nodistro main" > /etc/apt/sources.list.d/nodesource.list \
     && apt-get update \
     && apt-get install -y --no-install-recommends nodejs \
     && apt-get clean \

misc/completions/bash/_scratch

@@ -4,7 +4,7 @@ _shtab_scratch_subparsers=('login' 'create' 'list' 'ls' 'ssh' 'sftp' 'destroy' '
 
 _shtab_scratch_option_strings=('-h' '--help')
 _shtab_scratch_login_option_strings=('-h' '--help')
-_shtab_scratch_create_option_strings=('-h' '--help' '--key-name' '--security-group-name' '--extra-tags' '--instance-profile' '--output-format' '--git-rev' '--ssh' '--max-age-days')
+_shtab_scratch_create_option_strings=('-h' '--help' '--key-name' '--security-group-name' '--extra-tags' '--instance-profile' '--output-format' '--git-rev' '--ssh' '--instance-type' '--size-gb' '--max-age-days')
 _shtab_scratch_list_option_strings=('-h' '--help' '--all' '--output-format')
 _shtab_scratch_ls_option_strings=('-h' '--help' '--all' '--output-format')
 _shtab_scratch_ssh_option_strings=('-h' '--help')
@@ -14,8 +14,8 @@ _shtab_scratch_rm_option_strings=('-h' '--help' '--all-mine' '-y' '--yes' '--out
 _shtab_scratch_push_option_strings=('-h' '--help' '--rev')
 _shtab_scratch_completion_option_strings=('-h' '--help')
 _shtab_scratch_forward_option_strings=('-h' '--help')
-_shtab_scratch_go_option_strings=('-h' '--help' '--max-age-days')
-_shtab_scratch_claude_option_strings=('-h' '--help' '--max-age-days')
+_shtab_scratch_go_option_strings=('-h' '--help' '--instance-type' '--size-gb' '--max-age-days')
+_shtab_scratch_claude_option_strings=('-h' '--help' '--instance-type' '--size-gb' '--max-age-days')
 
 
 

misc/completions/zsh/_scratch

@@ -31,6 +31,8 @@ _shtab_scratch_defaults_added=0
 
 _shtab_scratch_claude_options=(
   "(- : *)"{-h,--help}"[show this help message and exit]"
+  "--instance-type[Override the EC2 instance type. The AMI is derived from its architecture. May be used without a machine config.]:instance_type:"
+  "--size-gb[Override the root volume size in GiB.]:size_gb:"
   "--max-age-days[Maximum age for scratch instance in days. Defaults to 1.5]:max_age_days:"
   ":Machine config name from misc\/scratch (e.g. dev-box).:"
 )
@@ -56,6 +58,8 @@ _shtab_scratch_create_options=(
   "--output-format[]:output_format:(table csv)"
   "--git-rev[Git revision of \`materialize\` codebase to push to scratch instance. Defaults to \`HEAD\`]:git_rev:"
   "--ssh[ssh into the machine after the launch script is run. Only works if a single instance was started]"
+  "--instance-type[Override the EC2 instance type. The AMI is derived from its architecture. May be used without a machine config.]:instance_type:"
+  "--size-gb[Override the root volume size in GiB.]:size_gb:"
   "--max-age-days[Maximum age for scratch instance in days. Defaults to 1.5]:max_age_days:"
   ":Use a config from \{machine\}.json in \`misc\/scratch\`. Hint\: \`dev-box\` is a good starter\!:"
 )
@@ -85,6 +89,8 @@ _shtab_scratch_forward_defaults_added=0
 
 _shtab_scratch_go_options=(
   "(- : *)"{-h,--help}"[show this help message and exit]"
+  "--instance-type[Override the EC2 instance type. The AMI is derived from its architecture. May be used without a machine config.]:instance_type:"
+  "--size-gb[Override the root volume size in GiB.]:size_gb:"
   "--max-age-days[Maximum age for scratch instance in days. Defaults to 1.5]:max_age_days:"
   ":Machine config name from misc\/scratch (e.g. dev-box). Used only when creating a new instance.:"
 )

misc/python/materialize/cli/scratch/claude.py

@@ -13,10 +13,12 @@
 import datetime
 
 from materialize.cli.scratch import check_required_vars, list_all_instances
-from materialize.cli.scratch.create import MAX_AGE_DAYS, multi_json, pick_machine
+from materialize.cli.scratch.create import (
+    MAX_AGE_DAYS,
+    load_machine_descs,
+    pick_machine,
+)
 from materialize.scratch import (
-    MZ_ROOT,
-    MachineDesc,
     launch_cluster,
     mssh,
     print_instances,
@@ -32,6 +34,19 @@ def configure_parser(parser: argparse.ArgumentParser) -> None:
         default=None,
         help="Machine config name from misc/scratch (e.g. dev-box).",
     )
+    parser.add_argument(
+        "--instance-type",
+        type=str,
+        help=(
+            "Override the EC2 instance type. The AMI is derived from its "
+            "architecture. May be used without a machine config."
+        ),
+    )
+    parser.add_argument(
+        "--size-gb",
+        type=int,
+        help="Override the root volume size in GiB.",
+    )
     parser.add_argument(
         "--max-age-days",
         type=float,
@@ -41,10 +56,18 @@ def configure_parser(parser: argparse.ArgumentParser) -> None:
 
 
 def run(args: argparse.Namespace) -> None:
-    machine = args.machine or pick_machine()
-
-    with open(MZ_ROOT / "misc" / "scratch" / f"{machine}.json") as f:
-        descs = [MachineDesc.model_validate(obj) for obj in multi_json(f.read())]
+    # Resolve the machine config name eagerly so it can be shown below, unless
+    # the user supplied a bare instance type instead.
+    machine = args.machine
+    if machine is None and args.instance_type is None:
+        machine = pick_machine()
+    label = machine or args.instance_type
+
+    descs = load_machine_descs(
+        machine,
+        instance_type=args.instance_type,
+        size_gb=args.size_gb,
+    )
 
     if len(descs) != 1:
         raise RuntimeError(
@@ -61,7 +84,7 @@ def run(args: argparse.Namespace) -> None:
 
     if active:
         print_instances(active, numbered=True)
-        print(f"  {len(active) + 1}) Create new instance ({machine})")
+        print(f"  {len(active) + 1}) Create new instance ({label})")
         while True:
             choice = input("Select an instance or create new [#]: ").strip()
             try:
@@ -79,13 +102,13 @@ def run(args: argparse.Namespace) -> None:
         instance = None
 
     if instance is None:
-        say(f"Creating from {machine}...")
+        say(f"Creating from {label}...")
         max_age = datetime.timedelta(days=args.max_age_days)
         extra_tags = {"LaunchedBy": whoami()}
         instances = launch_cluster(
             descs,
             extra_tags=extra_tags,
-            delete_after=datetime.datetime.utcnow() + max_age,
+            delete_after=datetime.datetime.now(datetime.timezone.utc) + max_age,
         )
 
         print("Launched:")

misc/python/materialize/cli/scratch/create.py

@@ -21,6 +21,7 @@
     launch_cluster,
     mssh,
     print_instances,
+    render_table,
     whoami,
 )
 
@@ -36,8 +37,6 @@ def _natsort_key(s: str) -> list:
 
 def pick_machine() -> str:
     """Show available machine configs and let the user pick one."""
-    from prettytable import PrettyTable
-
     scratch_dir = MZ_ROOT / "misc" / "scratch"
     configs: list[tuple[str, str, str]] = []
     for f in sorted(scratch_dir.glob("*.json"), key=lambda f: _natsort_key(f.stem)):
@@ -46,12 +45,13 @@ def pick_machine() -> str:
         for d in descs:
             configs.append((f.stem, d.instance_type, d.name))
 
-    pt = PrettyTable()
-    pt.field_names = ["#", "Config", "Instance Type", "Name"]
-    pt.align = "l"
-    for i, (stem, instance_type, name) in enumerate(configs, 1):
-        pt.add_row([i, stem, instance_type, name])
-    print(pt)
+    render_table(
+        ["#", "CONFIG", "INSTANCE TYPE", "NAME"],
+        [
+            [str(i), stem, instance_type, name]
+            for i, (stem, instance_type, name) in enumerate(configs, 1)
+        ],
+    )
 
     while True:
         choice = input("Select a machine config [#]: ").strip()
@@ -92,6 +92,44 @@ def multi_json(s: str) -> list[dict[Any, Any]]:
     return result
 
 
+def load_machine_descs(
+    machine: str | None,
+    *,
+    instance_type: str | None = None,
+    size_gb: int | None = None,
+) -> list[MachineDesc]:
+    """Load machine configs, applying optional overrides.
+
+    With no `machine` but an explicit `instance_type`, synthesize a single
+    default config so callers can launch an arbitrary instance type without a
+    preset file. With no `machine` and no `instance_type`, prompt for one."""
+    if machine is None and instance_type is not None:
+        desc = MachineDesc(name=instance_type, instance_type=instance_type)
+        if size_gb is not None:
+            desc.size_gb = size_gb
+        return [desc]
+
+    if machine is None:
+        machine = pick_machine()
+
+    with open(MZ_ROOT / "misc" / "scratch" / f"{machine}.json") as f:
+        print(f"Reading machine configs from {f.name}")
+        descs = [MachineDesc.model_validate(obj) for obj in multi_json(f.read())]
+
+    if instance_type is not None:
+        if len(descs) != 1:
+            raise RuntimeError(
+                f"--instance-type requires a single-machine config, got {len(descs)}"
+            )
+        descs[0].instance_type = instance_type
+        # An explicit type override invalidates any pinned AMI; re-derive it.
+        descs[0].ami = None
+    if size_gb is not None:
+        for d in descs:
+            d.size_gb = size_gb
+    return descs
+
+
 def configure_parser(parser: argparse.ArgumentParser) -> None:
     parser.add_argument(
         "--key-name", type=str, required=False, help="Optional EC2 Key Pair name"
@@ -139,6 +177,19 @@ def configure_parser(parser: argparse.ArgumentParser) -> None:
             "Hint: `dev-box` is a good starter!"
         ),
     )
+    parser.add_argument(
+        "--instance-type",
+        type=str,
+        help=(
+            "Override the EC2 instance type. The AMI is derived from its "
+            "architecture. May be used without a machine config."
+        ),
+    )
+    parser.add_argument(
+        "--size-gb",
+        type=int,
+        help="Override the root volume size in GiB.",
+    )
     parser.add_argument(
         "--max-age-days",
         type=float,
@@ -158,15 +209,11 @@ def run(args: argparse.Namespace) -> None:
                 "extra-tags must be a JSON dictionary of strings to strings"
             )
 
-    if args.machine:
-        with open(MZ_ROOT / "misc" / "scratch" / f"{args.machine}.json") as f:
-            print(f"Reading machine configs from {f.name}")
-            descs = [MachineDesc.model_validate(obj) for obj in multi_json(f.read())]
-    else:
-        args.machine = pick_machine()
-        with open(MZ_ROOT / "misc" / "scratch" / f"{args.machine}.json") as f:
-            print(f"Reading machine configs from {f.name}")
-            descs = [MachineDesc.model_validate(obj) for obj in multi_json(f.read())]
+    descs = load_machine_descs(
+        args.machine,
+        instance_type=args.instance_type,
+        size_gb=args.size_gb,
+    )
 
     check_required_vars()
     extra_tags["LaunchedBy"] = whoami()
@@ -184,7 +231,7 @@ def run(args: argparse.Namespace) -> None:
         security_group_name=args.security_group_name,
         instance_profile=args.instance_profile,
         extra_tags=extra_tags,
-        delete_after=datetime.datetime.utcnow() + max_age,
+        delete_after=datetime.datetime.now(datetime.timezone.utc) + max_age,
         git_rev=args.git_rev,
         extra_env={},
     )

misc/python/materialize/cli/scratch/go.py

@@ -14,10 +14,8 @@
 import datetime
 
 from materialize.cli.scratch import check_required_vars
-from materialize.cli.scratch.create import MAX_AGE_DAYS, multi_json
+from materialize.cli.scratch.create import MAX_AGE_DAYS, load_machine_descs
 from materialize.scratch import (
-    MZ_ROOT,
-    MachineDesc,
     get_instance,
     launch_cluster,
     mssh,
@@ -32,11 +30,26 @@
 def configure_parser(parser: argparse.ArgumentParser) -> None:
     parser.add_argument(
         "machine",
+        nargs="?",
+        default=None,
         help=(
             "Machine config name from misc/scratch (e.g. dev-box). "
             "Used only when creating a new instance."
         ),
     )
+    parser.add_argument(
+        "--instance-type",
+        type=str,
+        help=(
+            "Override the EC2 instance type. The AMI is derived from its "
+            "architecture. May be used without a machine config."
+        ),
+    )
+    parser.add_argument(
+        "--size-gb",
+        type=int,
+        help="Override the root volume size in GiB.",
+    )
     parser.add_argument(
         "--max-age-days",
         type=float,
@@ -46,9 +59,12 @@ def configure_parser(parser: argparse.ArgumentParser) -> None:
 
 
 def run(args: argparse.Namespace) -> None:
-    # Read the machine config
-    with open(MZ_ROOT / "misc" / "scratch" / f"{args.machine}.json") as f:
-        descs = [MachineDesc.model_validate(obj) for obj in multi_json(f.read())]
+    # Read the machine config, applying any overrides.
+    descs = load_machine_descs(
+        args.machine,
+        instance_type=args.instance_type,
+        size_gb=args.size_gb,
+    )
 
     if len(descs) != 1:
         raise RuntimeError(
@@ -69,14 +85,15 @@ def run(args: argparse.Namespace) -> None:
         pass
 
     # No existing instance — create one
-    say(f"No existing instance found, creating from {args.machine}...")
+    label = args.machine or args.instance_type
+    say(f"No existing instance found, creating from {label}...")
 
     max_age = datetime.timedelta(days=args.max_age_days)
     extra_tags = {"LaunchedBy": whoami()}
     instances = launch_cluster(
         descs,
         extra_tags=extra_tags,
-        delete_after=datetime.datetime.utcnow() + max_age,
+        delete_after=datetime.datetime.now(datetime.timezone.utc) + max_age,
     )
 
     print("Launched:")

misc/python/materialize/cli/scratch/list_cmd.py

@@ -25,4 +25,9 @@ def configure_parser(parser: argparse.ArgumentParser) -> None:
 
 def run(args: argparse.Namespace) -> None:
     instances = list_all_instances(owners=args.owner or None, all=args.all)
-    print_instances(instances, args.output_format)
+    # Show the owner column only when listing instances beyond your own.
+    print_instances(
+        instances,
+        args.output_format,
+        show_launched_by=args.all or bool(args.owner),
+    )