health check: add on-failure actions

For systems that have extreme robustness requirements (edge devices,
particularly those in difficult to access environments), it is important
that applications continue running in all circumstances. When the
application fails, Podman must restart it automatically to provide this
robustness. Otherwise, these devices may require customer IT to
physically gain access to restart, which can be prohibitively difficult.

Add a new `--on-failure` flag that supports four actions:

- **none**: Take no action.

- **kill**: Kill the container.

- **restart**: Restart the container.  Do not combine the `restart`
               action with the `--restart` flag.  When running inside of
               a systemd unit, consider using the `kill` or `stop`
               action instead to make use of systemd's restart policy.

- **stop**: Stop the container.

To remain backwards compatible, **none** is the default action.

Signed-off-by: Valentin Rothberg <vrothberg@redhat.com>

2 files changed, 76 insertions, 0 deletions

diff --git a/libpod/define/container_inspect.go b/libpod/define/container_inspect.go
index 5982d684c..da5c58f27 100644
--- a/libpod/define/container_inspect.go
+++ b/libpod/define/container_inspect.go
@@ -55,6 +55,8 @@ type InspectContainerConfig struct {
 	StopSignal uint `json:"StopSignal"`
 	// Configured healthcheck for the container
 	Healthcheck *manifest.Schema2HealthConfig `json:"Healthcheck,omitempty"`
+	// HealthcheckOnFailureAction defines an action to take once the container turns unhealthy.
+	HealthcheckOnFailureAction string `json:"HealthcheckOnFailureAction,omitempty"`
 	// CreateCommand is the full command plus arguments of the process the
 	// container has been created with.
 	CreateCommand []string `json:"CreateCommand,omitempty"`
diff --git a/libpod/define/healthchecks.go b/libpod/define/healthchecks.go
index f71274350..274e02561 100644
--- a/libpod/define/healthchecks.go
+++ b/libpod/define/healthchecks.go
@@ -1,5 +1,10 @@
 package define
 
+import (
+	"fmt"
+	"strings"
+)
+
 const (
 	// HealthCheckHealthy describes a healthy container
 	HealthCheckHealthy string = "healthy"
@@ -57,3 +62,72 @@ const (
 	// HealthConfigTestCmdShell runs commands with the system's default shell
 	HealthConfigTestCmdShell = "CMD-SHELL"
 )
+
+// HealthCheckOnFailureAction defines how Podman reacts when a container's health
+// status turns unhealthy.
+type HealthCheckOnFailureAction int
+
+// Healthcheck on-failure actions.
+const (
+	// HealthCheckOnFailureActionNonce instructs Podman to not react on an unhealthy status.
+	HealthCheckOnFailureActionNone = iota // Must be first iota for backwards compatibility
+	// HealthCheckOnFailureActionInvalid denotes an invalid on-failure policy.
+	HealthCheckOnFailureActionInvalid = iota
+	// HealthCheckOnFailureActionNonce instructs Podman to kill the container on an unhealthy status.
+	HealthCheckOnFailureActionKill = iota
+	// HealthCheckOnFailureActionNonce instructs Podman to restart the container on an unhealthy status.
+	HealthCheckOnFailureActionRestart = iota
+	// HealthCheckOnFailureActionNonce instructs Podman to stop the container on an unhealthy status.
+	HealthCheckOnFailureActionStop = iota
+)
+
+// String representations for on-failure actions.
+const (
+	strHealthCheckOnFailureActionNone    = "none"
+	strHealthCheckOnFailureActionInvalid = "invalid"
+	strHealthCheckOnFailureActionKill    = "kill"
+	strHealthCheckOnFailureActionRestart = "restart"
+	strHealthCheckOnFailureActionStop    = "stop"
+)
+
+// SupportedHealthCheckOnFailureActions lists all supported healthcheck restart policies.
+var SupportedHealthCheckOnFailureActions = []string{
+	strHealthCheckOnFailureActionNone,
+	strHealthCheckOnFailureActionKill,
+	strHealthCheckOnFailureActionRestart,
+	strHealthCheckOnFailureActionStop,
+}
+
+// String returns the string representation of the HealthCheckOnFailureAction.
+func (h HealthCheckOnFailureAction) String() string {
+	switch h {
+	case HealthCheckOnFailureActionNone:
+		return strHealthCheckOnFailureActionNone
+	case HealthCheckOnFailureActionKill:
+		return strHealthCheckOnFailureActionKill
+	case HealthCheckOnFailureActionRestart:
+		return strHealthCheckOnFailureActionRestart
+	case HealthCheckOnFailureActionStop:
+		return strHealthCheckOnFailureActionStop
+	default:
+		return strHealthCheckOnFailureActionInvalid
+	}
+}
+
+// ParseHealthCheckOnFailureAction parses the specified string into a HealthCheckOnFailureAction.
+// An error is returned for an invalid input.
+func ParseHealthCheckOnFailureAction(s string) (HealthCheckOnFailureAction, error) {
+	switch s {
+	case "", strHealthCheckOnFailureActionNone:
+		return HealthCheckOnFailureActionNone, nil
+	case strHealthCheckOnFailureActionKill:
+		return HealthCheckOnFailureActionKill, nil
+	case strHealthCheckOnFailureActionRestart:
+		return HealthCheckOnFailureActionRestart, nil
+	case strHealthCheckOnFailureActionStop:
+		return HealthCheckOnFailureActionStop, nil
+	default:
+		err := fmt.Errorf("invalid on-failure action %q for health check: supported actions are %s", s, strings.Join(SupportedHealthCheckOnFailureActions, ","))
+		return HealthCheckOnFailureActionInvalid, err
+	}
+}