This is a large squashed commit covering two batches of prior uncommitted work plus a full security-audit remediation pass, kept together because go.mod/go.sum and several shared files (main.go, handler.go) were touched by both and splitting risked non-building intermediate commits. Features (built earlier, previously uncommitted): - Local username/password login for single-tenant deployments with no SSO configured (api/localauth, alerting/internal/sessioncheck, sentryctl users, web/src/routes/login, metadata migrations 0040/0041). - Remotely-editable additional log file paths for agents, on top of their existing primary source (api/agents, agent/sentry-agent extra-file-path diffing, web agent config UI). - IPv4/IPv6 addresses reported alongside other host system metrics. Security audit remediation (this pass, all live-verified in production): - Critical: block ClickHouse SSRF table functions (url/remote/file/s3/...) in the raw-SQL query escape hatch. - High: deny sensitive paths and require Admin to add agent extra_file_paths (Editor could previously point an agent at /etc/shadow or an SSH key); alerting webhook targets now validate against internal/metadata/loopback addresses, both at creation and send time; alerting's session middleware now enforces an Editor+ floor on mutating requests instead of "any authenticated session"; bumped goxmldsig to close a SAML signature-verification bypass (GO-2026-4753). - Medium: per-IP login rate limiting; security response headers (HSTS/CSP/nosniff/X-Frame-Options/Referrer-Policy/Permissions-Policy) on web/nginx.conf; a DevCredentialWarnings check in every Go service's config loader, logging loudly at startup if a deployment is still on docker-compose.yml's literal dev-only credentials; dependency bumps (golang.org/x/text, grpc, x/net, quick-xml, h2) across every affected Go module and both Rust crates, including a previously-uncovered x/net vulnerability in deploy/operator; a new security-scan.yml CI workflow running cargo-deny/govulncheck/npm-audit, mirroring the existing license-compliance.yml matrix shape. - Low: removed sentryctl's plaintext --password flag (shell history/`ps` exposure) in favor of stdin and a --password-stdin flag for reset-password's optional specific-password path; a dummy bcrypt comparison closes a login response-time username-enumeration side-channel.
121 lines
5.8 KiB
Protocol Buffer
121 lines
5.8 KiB
Protocol Buffer
syntax = "proto3";
|
|
|
|
package sentry.agent.v1;
|
|
|
|
option go_package = "github.com/sentry/sentry/proto/sentry/agent/v1;agentv1";
|
|
|
|
// AgentControl is the control-plane counterpart to logs.v1.LogIngest's
|
|
// data-plane PushBatch -- the same mTLS channel/connection an agent
|
|
// already has open to ingest, a second gRPC service on the same
|
|
// listener rather than a second protocol or connection the agent would
|
|
// need to maintain (see /docs/agent-management-design.md). CheckIn is
|
|
// agent-initiated, called on the agent's own heartbeat ticker: there is
|
|
// still no path for the platform to reach into an agent uninvited. An
|
|
// agent asks "what should I be running" on its own schedule -- the same
|
|
// push-not-pull posture the heartbeat feature this builds on already
|
|
// established.
|
|
service AgentControl {
|
|
rpc CheckIn(CheckInRequest) returns (CheckInResponse);
|
|
}
|
|
|
|
// ReportedConfig is what an agent tells the platform about itself --
|
|
// read-only, for inventory/visibility. Deliberately excludes tls/ingest
|
|
// endpoint fields: those are never reported and never remotely
|
|
// overridable (see DesiredOverride's comment) -- reporting the ingest
|
|
// endpoint back to itself would be redundant (that's exactly the
|
|
// connection this request arrived over), and TLS material has no
|
|
// business leaving the host at all.
|
|
message ReportedConfig {
|
|
string agent_version = 1;
|
|
string source_kind = 2; // "journald", "file", "eventlog", "etw"
|
|
string source_detail = 3; // human-readable summary: unit name, file path, or channel list
|
|
uint64 batch_max_size = 4;
|
|
uint64 batch_flush_interval_ms = 5;
|
|
bool heartbeat_enabled = 6;
|
|
uint64 heartbeat_interval_ms = 7;
|
|
}
|
|
|
|
message CheckInRequest {
|
|
string host = 1;
|
|
string service = 2;
|
|
ReportedConfig current_config = 3;
|
|
// The DesiredOverride.version this agent last successfully applied,
|
|
// empty if it has never applied one. Lets the server distinguish
|
|
// "pending" (an edit exists the agent hasn't picked up yet) from
|
|
// "applied" for the web UI, without the agent needing to know
|
|
// anything about that distinction itself.
|
|
string applied_override_version = 4;
|
|
}
|
|
|
|
// DesiredOverride is the remotely-editable subset of an agent's config
|
|
// -- batch/heartbeat tuning, and, for journald sources, the unit
|
|
// filter. Every field is optional: unset means "no override for this
|
|
// field, keep whatever agent.toml says locally" -- a partial edit only
|
|
// touches the fields it sets. Never includes tls/ingest: those stay
|
|
// local-file-only, permanently, a deliberate security boundary (see
|
|
// /docs/agent-management-design.md) so a bad or malicious remote edit
|
|
// can never strand an agent or redirect where its logs go.
|
|
message DesiredOverride {
|
|
optional uint64 batch_max_size = 1;
|
|
optional uint64 batch_flush_interval_ms = 2;
|
|
optional bool heartbeat_enabled = 3;
|
|
optional uint64 heartbeat_interval_ms = 4;
|
|
// Only meaningful when the agent's local source is journald; ignored
|
|
// otherwise. Empty string means "no unit filter" (tail the whole
|
|
// journal), same semantics as the local config's own unit field.
|
|
optional string journald_unit = 5;
|
|
// Opaque version stamp the platform assigns on every edit. The
|
|
// agent's only obligation is to echo it back as
|
|
// CheckInRequest.applied_override_version once applied -- it never
|
|
// interprets the value itself.
|
|
string version = 6;
|
|
// Extra file paths this agent should tail in addition to whatever its
|
|
// local [source] already is -- never a replacement for the primary
|
|
// source (an agent whose local source is journald can still be told
|
|
// to also tail a file, and vice versa). Unlike every field above,
|
|
// there's no real "unset" state for a list: the web UI/CLI always
|
|
// resubmit the complete desired list on every edit (same "PUT
|
|
// replaces the whole override" convention every other field already
|
|
// follows -- see api/agents/handler.go's handleSetConfig), so an
|
|
// empty list unambiguously means "no extra paths right now," not
|
|
// "don't touch this."
|
|
repeated string extra_file_paths = 7;
|
|
}
|
|
|
|
// AgentCommand is a one-shot action, not a persistent desired state like
|
|
// DesiredOverride -- delivered at-most-once (see CheckInResponse's
|
|
// comment). Scoped deliberately narrow: only RESTART exists today.
|
|
// STOP and UNINSTALL are real, disclosed future work, not oversights --
|
|
// both need genuine OS service-manager integration (systemd's
|
|
// Restart=/RestartPreventExitStatus= semantics vs. Windows SCM recovery
|
|
// options are different enough per platform that hand-waving them would
|
|
// be dishonest), which RESTART doesn't: a graceful shutdown followed by
|
|
// a clean process exit, relying on whatever restart policy the host's
|
|
// service manager already has configured -- the same contract systemd/
|
|
// SCM already expect from any well-behaved service.
|
|
enum AgentCommand {
|
|
AGENT_COMMAND_UNSPECIFIED = 0;
|
|
AGENT_COMMAND_RESTART = 1;
|
|
}
|
|
|
|
message CheckInResponse {
|
|
// False when no override has ever been set for this agent -- it
|
|
// should be running whatever agent.toml already has, untouched.
|
|
bool has_override = 1;
|
|
DesiredOverride override = 2;
|
|
// AGENT_COMMAND_UNSPECIFIED when there's nothing to do. Unlike
|
|
// DesiredOverride, there is no "applied_command_version" echoed back
|
|
// in CheckInRequest: the server clears a pending command the moment
|
|
// it hands it out in a response (see
|
|
// ingest/internal/agentregistry.Registry.CheckIn), not once the agent
|
|
// confirms execution -- a restarting agent's process is gone before
|
|
// it could ever send that confirmation. This is an honest at-most-
|
|
// once delivery, not at-least-once: a command lost to a network
|
|
// failure between this response and the agent acting on it is simply
|
|
// lost, same as any fire-and-forget signal. Re-issuing (PUT
|
|
// /agents/{host}/command again) is the operator's recourse, same as
|
|
// it would be for a `systemctl restart` that silently failed to reach
|
|
// its target.
|
|
AgentCommand pending_command = 3;
|
|
}
|