Fix netdatacli ping command (#21965)
* Refactor API endpoints to use `netdata_ready_load()` and enhance `ping` command to check agent initialization status. * Adjust reply prefix for ping command * Update src/cli/README.md Co-authored-by: Ilya Mashchenko <ilya@netdata.cloud> * Address review comments * Potential fix for pull request finding Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> * Cleanup description * Update src/cli/README.md Co-authored-by: cubic-dev-ai[bot] <191113872+cubic-dev-ai[bot]@users.noreply.github.com> --------- Co-authored-by: Ilya Mashchenko <ilya@netdata.cloud> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> Co-authored-by: cubic-dev-ai[bot] <191113872+cubic-dev-ai[bot]@users.noreply.github.com>
Stelios Fragkakis committed
Mar 19, 2026 at 09:11 UTC
bb84d228173e55075cb0292480b74a4146419546
11 files changed
+31
-15
src/cli/README.md
+2
-2
@@ -10,12 +10,12 @@ Available commands:
10
|---------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
11
| `help` | Display usage information and exit. |
12
| `reload-health` | Reloads the Netdata health configuration, updating alerts based on changes made to configuration files. |
13
-| `reload-labels` | Reloads [host labels](/docs/netdata-agent/configuration/organize-systems-metrics-and-alerts.md#create-custom-labels) from netdata.conf. |
13
+| `reload-labels` | Reloads [host labels](/docs/netdata-agent/configuration/organize-systems-metrics-and-alerts.md#create-custom-labels) from netdata.conf. |
14
| `reopen-logs` | Close and reopen log files. |
15
| `shutdown-agent` | Gracefully shut down the Netdata Agent. |
16
| `fatal-agent` | Log the current state and forcefully halt the Netdata Agent. |
17
| `reload-claiming-state` | Reload the Agent's claiming state from disk. |
18
-| `ping` | Checks the Agent's status. If the Agent is alive, it exits with status code 0 and prints 'pong' to standard output. Exits with status code 255 otherwise. |
18
+| `ping` | Verifies whether the Agent is ready. Outputs `pong` when the Agent command server can be contacted. Exit code `0` when the Agent is ready to receive commands. Exit code `1` while the Agent is still initializing. Exit code `255` if the Agent is unreachable |
19
| `aclk-state [json]` | Return the current state of ACLK and Cloud connection. Optionally in JSON. |
20
| `dumpconfig` | Display the current netdata.conf configuration. |
21
| `mark-stale-nodes-ephemeral <node_id \| machine_guid \| hostname \| ALL_NODES>` | Marks one or all disconnected nodes, including virtual nodes, as [ephemeral](/docs/nodes-ephemerality.md), while keeping their previously collected metrics data available for queries on both this Netdata Agent dashboard and Netdata Cloud.[^1][^2] |
src/daemon/commands.c
+12
-3
@@ -33,6 +33,14 @@ struct command_context {
33
unsigned command_string_size;
34
};
35
36
+static inline char command_reply_prefix(const struct command_context *cmd_ctx, cmd_status_t status)
37
+{
38
+ if (cmd_ctx->idx == CMD_PING)
39
+ return CMD_PREFIX_INFO;
40
+
41
+ return cmd_prefix_by_status[status];
42
+}
43
+
44
/* Forward declarations */
45
static cmd_status_t cmd_help_execute(char *args, char **message);
46
static cmd_status_t cmd_reload_health_execute(char *args, char **message);
@@ -61,7 +69,7 @@ static command_info_t command_info_array[] = {
69
{"reload-labels", "", "Reload all localhost labels.", cmd_reload_labels_execute, CMD_TYPE_ORTHOGONAL, CMD_INIT_STATUS_FULL}, // reload the labels
70
{"read-config", "", "", cmd_read_config_execute, CMD_TYPE_CONCURRENT, CMD_INIT_STATUS_FULL},
71
{"write-config", "", "", cmd_write_config_execute, CMD_TYPE_ORTHOGONAL, CMD_INIT_STATUS_FULL},
64
- {"ping", "", "Return with 'pong' if agent is alive.", cmd_ping_execute, CMD_TYPE_ORTHOGONAL, CMD_INIT_STATUS_INIT}, // ping command
72
+ {"ping", "", "Return with 'pong'; exit 0 when ready, 1 while initializing, 255 if the agent cannot be contacted.", cmd_ping_execute, CMD_TYPE_ORTHOGONAL, CMD_INIT_STATUS_INIT}, // ping command
73
{"aclk-state", "[json]", "Returns current state of ACLK and Netdata Cloud connection. (optionally in json).", cmd_aclk_state, CMD_TYPE_ORTHOGONAL, CMD_INIT_STATUS_FULL},
74
{"version", "", "Returns the netdata version.", cmd_version, CMD_TYPE_ORTHOGONAL, CMD_INIT_STATUS_INIT},
75
{"dumpconfig", "", "Returns the current netdata.conf on stdout.", cmd_dumpconfig, CMD_TYPE_ORTHOGONAL, CMD_INIT_STATUS_FULL},
@@ -309,7 +317,7 @@ static cmd_status_t cmd_ping_execute(char *args, char **message)
317
318
*message = strdupz("pong");
319
312
- return CMD_STATUS_SUCCESS;
320
+ return netdata_ready_load() ? CMD_STATUS_SUCCESS : CMD_STATUS_FAILURE;
321
}
322
323
static cmd_status_t cmd_aclk_state(char *args, char **message)
@@ -585,7 +593,7 @@ static void send_command_reply(struct command_context *cmd_ctx, cmd_status_t sta
593
add_char_to_command_reply(reply_string, &reply_string_size, '\0');
594
595
if (message) {
588
- add_char_to_command_reply(reply_string, &reply_string_size, cmd_prefix_by_status[status]);
596
+ add_char_to_command_reply(reply_string, &reply_string_size, command_reply_prefix(cmd_ctx, status));
597
add_string_to_command_reply(reply_string, &reply_string_size, message);
598
}
599
@@ -727,6 +735,7 @@ static void connection_cb(uv_stream_t *server, int status)
735
736
/* combined allocation of client pipe and command context */
737
cmd_ctx = mallocz(sizeof(*cmd_ctx));
738
+ cmd_ctx->idx = CMD_HELP;
739
client = (uv_pipe_t *)cmd_ctx;
740
ret = uv_pipe_init(server->loop, client, 1);
741
if (ret) {
src/daemon/common.h
+8
@@ -101,6 +101,14 @@ void system_tz_free(SYSTEM_TZ *tz);
101
extern bool netdata_ready;
102
extern time_t netdata_start_time;
103
104
+static inline bool netdata_ready_load(void) {
105
+ return __atomic_load_n(&netdata_ready, __ATOMIC_ACQUIRE);
106
+}
107
+
108
+static inline void netdata_ready_store(bool ready) {
109
+ __atomic_store_n(&netdata_ready, ready, __ATOMIC_RELEASE);
110
+}
111
+
112
void set_environment_for_plugins_and_scripts(void);
113
114
#ifdef __cplusplus
src/daemon/main.c
+2
-2
@@ -277,7 +277,7 @@ int netdata_main(int argc, char **argv) {
277
278
static_threads = static_threads_get();
279
280
- netdata_ready = false;
280
+ netdata_ready_store(false);
281
// set the name for logging
282
program_name = "netdata";
283
@@ -1156,7 +1156,7 @@ int netdata_main(int argc, char **argv) {
1156
(ready_ut - started_ut) / USEC_PER_MS, median_start_time / USEC_PER_MS);
1157
1158
cleanup_agent_event_log();
1159
- netdata_ready = true;
1159
+ netdata_ready_store(true);
1160
1161
// ----------------------------------------------------------------------------------------------------------------
1162
src/web/api/v1/api_v1_aclk.c
+1
-2
@@ -5,7 +5,7 @@
5
int api_v1_aclk(RRDHOST *host, struct web_client *w, char *url) {
6
UNUSED(url);
7
UNUSED(host);
8
- if (!netdata_ready) return HTTP_RESP_SERVICE_UNAVAILABLE;
8
+ if (!netdata_ready_load()) return HTTP_RESP_SERVICE_UNAVAILABLE;
9
10
BUFFER *wb = w->response.data;
11
buffer_flush(wb);
@@ -17,4 +17,3 @@ int api_v1_aclk(RRDHOST *host, struct web_client *w, char *url) {
17
buffer_no_cacheable(wb);
18
return HTTP_RESP_OK;
19
}
20
-
src/web/api/v1/api_v1_dbengine.c
+1
-1
@@ -71,7 +71,7 @@ static void web_client_api_v1_dbengine_stats_for_tier(BUFFER *wb, size_t tier) {
71
}
72
73
int api_v1_dbengine_stats(RRDHOST *host __maybe_unused, struct web_client *w, char *url __maybe_unused) {
74
- if (!netdata_ready)
74
+ if (!netdata_ready_load())
75
return HTTP_RESP_SERVICE_UNAVAILABLE;
76
77
BUFFER *wb = w->response.data;
src/web/api/v1/api_v1_function.c
+1
-1
@@ -3,7 +3,7 @@
3
#include "api_v1_calls.h"
4
5
int api_v1_function(RRDHOST *host, struct web_client *w, char *url) {
6
- if (!netdata_ready)
6
+ if (!netdata_ready_load())
7
return HTTP_RESP_SERVICE_UNAVAILABLE;
8
9
int timeout = 0;
src/web/api/v1/api_v1_functions.c
+1
-1
@@ -3,7 +3,7 @@
3
#include "api_v1_calls.h"
4
5
int api_v1_functions(RRDHOST *host, struct web_client *w, char *url __maybe_unused) {
6
- if (!netdata_ready)
6
+ if (!netdata_ready_load())
7
return HTTP_RESP_SERVICE_UNAVAILABLE;
8
9
BUFFER *wb = w->response.data;
src/web/api/v1/api_v1_info.c
+1
-1
@@ -165,7 +165,7 @@ static int web_client_api_request_v1_info_fill_buffer(RRDHOST *host, BUFFER *wb)
165
166
int api_v1_info(RRDHOST *host, struct web_client *w, char *url) {
167
(void)url;
168
- if (!netdata_ready) return HTTP_RESP_SERVICE_UNAVAILABLE;
168
+ if (!netdata_ready_load()) return HTTP_RESP_SERVICE_UNAVAILABLE;
169
BUFFER *wb = w->response.data;
170
buffer_flush(wb);
171
wb->content_type = CT_APPLICATION_JSON;
src/web/api/v1/api_v1_ml_info.c
+1
-1
@@ -6,7 +6,7 @@ int api_v1_ml_info(RRDHOST *host, struct web_client *w, char *url) {
6
(void) url;
7
#if defined(ENABLE_ML)
8
9
- if (!netdata_ready)
9
+ if (!netdata_ready_load())
10
return HTTP_RESP_SERVICE_UNAVAILABLE;
11
12
BUFFER *wb = w->response.data;
src/web/api/v2/api_v2_weights.c
+1
-1
@@ -3,7 +3,7 @@
3
#include "api_v2_calls.h"
4
5
int web_client_api_request_weights(RRDHOST *host, struct web_client *w, char *url, WEIGHTS_METHOD method, WEIGHTS_FORMAT format, size_t api_version) {
6
- if (!netdata_ready)
6
+ if (!netdata_ready_load())
7
return HTTP_RESP_SERVICE_UNAVAILABLE;
8
9
time_t baseline_after = 0, baseline_before = 0, after = 0, before = 0;